Pulse field ablation-based myocardial ablation catheter and control system

Abstract

Provided are a pulse field ablation-based myocardial ablation catheter and a control system. The system consists of a data acquisition module, a real-time data processing module, a nonlinear model calculation module, a control algorithm module, an electric field control module, a contact force adjustment module, and a safety protection module. By means of the combination of the output of a PID controller, the optimal parameter from Bayesian optimization, and the result of contact force adjustment, a final control signal is generated. By means of the integration of multiple control strategies, it is ensured that electric field control is not only highly responsive but also precise and accurate.

Description

A myocardial ablation catheter and control system based on pulsed field ablation Technical Field

[0001] This invention relates to the field of radiofrequency ablation technology, specifically to a myocardial ablation catheter and control system based on pulsed field ablation. Background Technology

[0002] Atrial fibrillation (AF) is a common cardiac arrhythmia. Current treatments for AF primarily include medication and catheter ablation. Ablation involves using heat or cold to coagulate and necrotize the pulmonary veins and surrounding tissues, thereby isolating the pulmonary veins and terminating the AF treatment.

[0003] The prior art CN 118766578 A describes a catheter based on pulsed field ablation, which has an adjustment system connected to the electrode assembly. The adjustment system is used to control the electric field generated by the electrode assembly, and has the effect of real-time adjustment of the ablation catheter and electrode during treatment. Technical issues

[0004] However, pulsed field ablation requires discharge to be completed at the microsecond level, and the adjustment system needs to respond to changes in parameters such as impedance and adhesion force in real time. The response speed of the adjustment system in the aforementioned technical solution is insufficient, which may lead to inaccurate control of the ablation depth and affect the treatment effect. The aforementioned technical solution uses a regression model for calculation, but the relationship between the ablation depth and discharge parameters such as voltage, pulse width and number of times in pulsed field ablation is non-linear. Forcing the use of a linear regression model will result in poor model fitting and inaccurate prediction results. Technical solutions

[0005] To address the shortcomings of existing technologies, this invention provides a myocardial ablation catheter and control system based on pulsed field ablation, which has the advantages of fast and precise electric field control for adjusting pulsed field ablation, and accurate fitting and prediction results of ablation lesion depth and discharge parameters, thus solving the problems of the aforementioned technologies.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a myocardial ablation catheter based on pulse field ablation, comprising a catheter, wherein the catheter is also connected to a pulse generating body for driving itself to generate an electric pulse field, the catheter having a built-in sensor for acquiring discharge parameters, the discharge parameters including voltage, pulse width and number of pulses, the sensor transmitting the acquired discharge parameters to a chip built into the pulse generating body via wires for modeling and calculation to obtain the optimal electric field distribution.

[0007] A myocardial ablation catheter control system based on pulsed field ablation, the system comprising a data acquisition module, a real-time data processing module, a nonlinear model calculation module, a control algorithm module, an electric field control module, a contact force adjustment module, and a safety protection module;

[0008] The data acquisition module is used to acquire impedance, contact force and electric field distribution data in real time, connects to the electric field control module and the contact force sensor, and performs filtering and standardization processing on the acquired data.

[0009] The real-time data processing module is used to process real-time data using an FPGA / ASIC accelerator, calculate the rate of change of impedance and the rate of change of contact force, and provide data to the control algorithm and model calculation module.

[0010] The nonlinear model calculation module is used to calculate the depth of ablation lesions using a neural network machine learning model, while dynamically updating model parameters, supporting online learning, and providing prediction results of the depth of ablation lesions.

[0011] The control algorithm module is used to implement a PID controller, dynamically adjust the discharge parameters, optimize the combination of discharge parameters using a Bayesian optimization algorithm, and provide control signals to the electric field control module.

[0012] The electric field control module is used to receive control signals, adjust the discharge parameters of the high-frequency pulse generator, adjust the impedance matching circuit in real time, optimize the electric field distribution, control the pulse discharge process, and ensure precise electric field control.

[0013] The contact force adjustment module is used to dynamically adjust the contact state between the catheter and myocardial tissue based on the contact force sensor data, so as to ensure that the catheter and tissue are in full contact and optimize the electric field distribution.

[0014] The safety protection module is used to monitor voltage, current and impedance parameters in real time, detect abnormal situations such as overvoltage, overcurrent and impedance sudden changes and trigger the protection mechanism, and provide safety alarms and handling suggestions.

[0015] Preferably, the filtering calculation of the data acquisition module is as follows:

[0016] ;

[0017] in: The filtered output at the current moment, This is the filtered output from the previous time step. The input data value at the current moment. Indicates the current input value Contribution to the current filtered output;

[0018] The data standardization processing calculation is as follows:

[0019]

[0020] in: This is the original data. The standard deviation of the data; The mean of the data. These are the standardized data values.

[0021] Preferably, the impedance change rate is calculated in the real-time data processing module as follows:

[0022]

[0023] Where: the rate of change of impedance is ; It is the impedance value at the current moment. It is the impedance value at the previous moment. It is a time interval, in seconds.

[0024] Preferably, the rate of change of contact force in the real-time data processing module is calculated as follows:

[0025]

[0026] Among them: the rate of change of the contact force ; It is the current contact force value. It is the contact force value at the previous moment.

[0027] Preferably, the neural network model in the nonlinear model calculation module is:

[0028] Assuming the neural network model is a multilayer perceptron (MLP), and the ablation lesion depth is... The prediction formula is:

[0029]

[0030] in: It is the input feature vector at the current moment. It is the weight matrix of the neural network. It is the bias vector of the neural network. It is a non-linear function of a neural network.

[0031] Preferably, the online learning update calculation in the nonlinear model calculation module is as follows:

[0032] During online learning, model parameters and Update dynamically based on new data, using gradient descent:

[0033]

[0034] in: It is the mean squared error of the loss function. It's the learning rate. and These are the gradients of the loss function with respect to the weights and biases, respectively. For the updated model parameters , For the updated model parameters .

[0035] Preferably, the depth of the ablation lesion is calculated in the nonlinear model calculation module as follows:

[0036] Obtain input feature data from the real-time data processing module. :

[0037]

[0038] Predict the depth of the ablation lesion using the current model parameters:

[0039]

[0040] Based on the new input data and the actual ablation depth Update model parameters:

[0041]

[0042]

[0043] Predicted ablation depth Provided to the control algorithm or model calculation module.

[0044] Preferably, the PID controller in the control algorithm module calculates as follows:

[0045]

[0046] in: For control signals; For proportional gain; For integral gain; This is the differential gain; This is the error signal, which is the difference between the set value and the actual value;

[0047] The Bayesian optimization algorithm calculates as follows:

[0048]

[0049] in: This is a combination of discharge parameters; It is the mean vector; It is the covariance matrix;

[0050] The Bayesian optimization algorithm also includes a posterior distribution, which is calculated as follows:

[0051]

[0052] in: These are the observed values; The observation matrix; Here is the noise covariance matrix;

[0053] The optimal combination of discharge parameters is calculated as follows:

[0054]

[0055] in: The optimal combination of discharge parameters is obtained through Bayesian optimization; Indicates parameters It is a parameter vector;

[0056] The control signal generation calculation is as follows:

[0057]

[0058] in: To provide control signals to the electric field control module; To be based on the optimal combination of discharge parameters The control signals for calculation.

[0059] Preferably, the contact force adjustment calculation in the contact force adjustment module is as follows:

[0060]

[0061] in: For the desired fit; The contact force detected by the sensor in real time; To dynamically adjust the gain and optimize the contact state between the catheter and myocardial tissue;

[0062] The dynamically adjusted gain is calculated as follows:

[0063]

[0064] in: This is the contact force adjustment coefficient; Adhesion to the target

[0065] The optimal electric field distribution is calculated as follows:

[0066]

[0067] in: The optimized electric field distribution; Let be the electric field distribution function. Beneficial effects

[0068] Compared with the prior art, the present invention provides a myocardial ablation catheter and control system based on pulsed field ablation, which has the following beneficial effects:

[0069] 1. This invention utilizes a PID controller to dynamically adjust the control signal, enabling rapid response to system error changes. The PID controller quickly adjusts the system output when errors occur, ensuring rapid response of the electric field control. A Bayesian optimization algorithm is used to optimize the discharge parameter combination, ensuring precise adjustment of the electric field distribution. Parameter optimization is guided by maximizing resource utilization and minimizing error, avoiding local optima. Bayesian optimization searches for the optimal parameter combination globally, achieving precise adjustment of the electric field control. The contact state between the catheter and myocardial tissue is dynamically adjusted based on sensor data, ensuring sufficient contact. Dynamic adjustment of the contact force further ensures sufficient contact between the catheter and myocardial tissue, optimizing the electric field distribution and avoiding control deviations caused by poor contact. The output of the PID controller, the optimal parameters from Bayesian optimization, and the results of contact force adjustment are combined to generate the final control signal. By integrating multiple control strategies, the electric field control is ensured to be not only fast-responding but also accurate, achieving the beneficial effects of fast response and precise control in pulsed field ablation.

[0070] 2. This invention evaluates the model's predictive accuracy by using mean squared error and coefficient of determination, employs cross-validation to ensure the model's generalization ability, and uses optimization algorithms such as gradient descent and genetic algorithms to optimize model parameters and improve fitting accuracy. It also seeks the optimal hyperparameter combination through grid search and random search. During the ablation process, discharge parameters are collected in real time, and the trained model is used to predict the ablation lesion depth. The predicted ablation lesion depth is compared with the target depth, and the ablation lesion depth is adjusted by controlling the discharge parameters, achieving accurate fitting and prediction. This invention achieves the beneficial effect of accurately predicting the ablation lesion depth and discharge parameters. Attached Figure Description

[0071] Figure 1 is a schematic diagram of the system of the present invention. Embodiments of the present invention

[0072] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0073] Please refer to Figure 1. A myocardial ablation catheter based on pulse field ablation includes a catheter and a pulse generator connected to drive itself to generate an electric pulse field. The catheter has a built-in sensor for acquiring discharge parameters, including voltage, pulse width, and number of pulses. The sensor transmits the acquired discharge parameters to a chip built into the pulse generator via wires for modeling and calculation to obtain the optimal electric field distribution.

[0074] A myocardial ablation catheter control system based on pulsed field ablation, the system comprising a data acquisition module, a real-time data processing module, a nonlinear model calculation module, a control algorithm module, an electric field control module, a contact force adjustment module, and a safety protection module;

[0075] The data acquisition module is used to acquire impedance, contact force and electric field distribution data in real time, connects to the electric field control module and the contact force sensor, and performs filtering and standardization processing on the acquired data.

[0076] The real-time data processing module is used to process real-time data using an FPGA / ASIC accelerator, calculate the rate of change of impedance and the rate of change of contact force, and provide data to the control algorithm and model calculation module.

[0077] The nonlinear model calculation module is used to calculate the depth of ablation lesions using a neural network machine learning model, while dynamically updating model parameters, supporting online learning, and providing prediction results of the depth of ablation lesions.

[0078] The control algorithm module is used to implement a PID controller, dynamically adjust the discharge parameters, optimize the combination of discharge parameters using a Bayesian optimization algorithm, and provide control signals to the electric field control module.

[0079] The electric field control module is used to receive control signals, adjust the discharge parameters of the high-frequency pulse generator, adjust the impedance matching circuit in real time, optimize the electric field distribution, control the pulse discharge process, and ensure precise electric field control.

[0080] The contact force adjustment module is used to dynamically adjust the contact state between the catheter and myocardial tissue based on the contact force sensor data, so as to ensure that the catheter and tissue are in full contact and optimize the electric field distribution.

[0081] The safety protection module is used to monitor voltage, current and impedance parameters in real time, detect abnormal situations such as overvoltage, overcurrent and impedance sudden changes and trigger the protection mechanism, and provide safety alarms and handling suggestions.

[0082] PID controllers can quickly respond to changes in system error by dynamically adjusting the control signal.

[0083] By properly setting the proportional gain, integral gain, and derivative gain, the discharge parameters can be quickly adjusted.

[0084] PID controllers can quickly adjust the system output when errors occur, ensuring a fast response in electric field control.

[0085] Bayesian optimization algorithms are used to optimize the combination of discharge parameters to ensure precise adjustment of the electric field distribution.

[0086] This algorithm guides parameter optimization by maximizing resource utilization and minimizing error, thus avoiding local optima.

[0087] Bayesian optimization can search for the optimal combination of parameters globally, enabling precise adjustment of electric field control.

[0088] The contact state between the catheter and myocardial tissue is dynamically adjusted based on sensor data to ensure sufficient contact between the catheter and myocardial tissue.

[0089] The gain coefficient of the contact force is dynamically adjusted to ensure a stable contact state.

[0090] By dynamically adjusting the contact force, the catheter is ensured to make full contact with the myocardial tissue, the electric field distribution is optimized, and control deviations caused by poor contact are avoided.

[0091] The final control signal is generated by combining the output of the PID controller, the optimal parameters obtained through Bayesian optimization, and the results of the contact force adjustment.

[0092] By integrating multiple control strategies, we ensure that electric field control is not only fast-responding but also accurate.

[0093] Preferably, the filtering calculation of the data acquisition module is as follows:

[0094]

[0095] in: The filtered output at the current moment, This is the filtered output from the previous time step. The input data value at the current moment. Indicates the current input value Contribution to the current filtered output;

[0096] The data standardization processing calculation is as follows:

[0097]

[0098] in: The original data, The standard deviation of the data; The mean of the data. These are the standardized data values.

[0099] Preferably, the impedance change rate is calculated in the real-time data processing module as follows:

[0100]

[0101] Where: the rate of change of impedance is ; It is the impedance value at the current moment. It is the impedance value at the previous moment. It is a time interval, in seconds.

[0102] Preferably, the rate of change of contact force in the real-time data processing module is calculated as follows:

[0103]

[0104] Among them: the rate of change of the contact force ; It is the current contact force value. It is the contact force value at the previous moment.

[0105] Preferably, the neural network model in the nonlinear model calculation module is:

[0106] Assuming the neural network model is a multilayer perceptron (MLP), the prediction formula for the ablation foci depth \( D(t) \) is:

[0107]

[0108] in: It is the input feature vector at the current moment. It is the weight matrix of the neural network. It is the bias vector of the neural network. It is a non-linear function of a neural network.

[0109] Preferably, the online learning update calculation in the nonlinear model calculation module is as follows:

[0110] During online learning, model parameters and Dynamically updated based on new data, the update is based on gradient descent:

[0111]

[0112]

[0113] in: It is the mean squared error of the loss function. It's the learning rate. and These are the gradients of the loss function with respect to the weights and biases, respectively.

[0114] Preferably, the depth of the ablation lesion is calculated in the nonlinear model calculation module as follows:

[0115] Obtain input feature data from the real-time data processing module. :

[0116]

[0117] Predict the depth of the ablation lesion using the current model parameters:

[0118]

[0119] Based on the new input data and the actual ablation depth Update model parameters:

[0120]

[0121]

[0122] Predicted ablation depth Provided to the control algorithm or model calculation module.

[0123] Preferably, the PID controller in the control algorithm module calculates as follows:

[0124]

[0125] in: For control signals; For proportional gain; For integral gain; This is the differential gain; This is the error signal, which is the difference between the set value and the actual value;

[0126] The Bayesian optimization algorithm calculates as follows:

[0127]

[0128] in: This is a combination of discharge parameters; It is the mean vector; It is the covariance matrix;

[0129] The Bayesian optimization algorithm also includes a posterior distribution, which is calculated as follows:

[0130]

[0131] in: These are the observed values; The observation matrix; Here is the noise covariance matrix;

[0132] The optimal combination of discharge parameters is calculated as follows:

[0133]

[0134] in: The optimal combination of discharge parameters is obtained through Bayesian optimization; Indicates parameters It is a parameter vector;

[0135] The control signal generation calculation is as follows:

[0136]

[0137] in: To provide control signals to the electric field control module; To be based on the optimal combination of discharge parameters The control signals for calculation.

[0138] Preferably, the contact force adjustment calculation in the contact force adjustment module is as follows:

[0139]

[0140] in: For the desired fit; The contact force detected by the sensor in real time; To dynamically adjust the gain and optimize the contact state between the catheter and myocardial tissue;

[0141] The dynamically adjusted gain is calculated as follows:

[0142]

[0143] in: This is the contact force adjustment coefficient; Adhesion to the target

[0144] The optimal electric field distribution is calculated as follows:

[0145]

[0146] in: The optimized electric field distribution; Let be the electric field distribution function.

[0147] Furthermore, by collecting discharge parameters (such as voltage, current, and frequency) and corresponding ablation lesion depth data during the ablation process, the collected data undergoes preprocessing steps such as cleaning and normalization to improve the model's training effect. A suitable machine learning or deep learning model is then selected for fitting and prediction. For example, regression models or neural networks can be used. The model is trained using the preprocessed data, and the fitting effect is optimized by adjusting the model parameters. Appropriate evaluation metrics (such as mean squared error and coefficient of determination) are used to assess the model's predictive accuracy. Cross-validation and other methods are employed to ensure the model's generalization ability. Optimization algorithms (such as gradient descent and genetic algorithms) are used to optimize model parameters and improve fitting accuracy. Optimal hyperparameter combinations are found through grid search and random search methods. During the ablation process, discharge parameters are collected in real time, and the trained model is used to predict the ablation lesion depth. The predicted ablation lesion depth is compared with the target depth, and the ablation lesion depth is adjusted by controlling the discharge parameters to achieve accurate fitting and prediction.

[0148] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A myocardial ablation catheter based on pulsed field ablation, comprising a catheter, wherein the catheter is further connected to a pulse generating body for driving itself to generate an electrical pulse field, characterized in that: The catheter has a built-in sensor for acquiring discharge parameters, including voltage, pulse width, and number of pulses. The sensor transmits the acquired discharge parameters to a chip built into the pulse generator via wires for modeling and calculation to obtain the optimal electric field distribution.

2. A myocardial ablation catheter control system based on pulsed field ablation, characterized in that, The control system is used in the myocardial ablation catheter of claim 1. The system consists of a data acquisition module, a real-time data processing module, a nonlinear model calculation module, a control algorithm module, an electric field control module, a contact force adjustment module, and a safety protection module. The data acquisition module is used to acquire impedance, contact force and electric field distribution data in real time, connects to the electric field control module and the contact force sensor, and performs filtering and standardization processing on the acquired data. The real-time data processing module is used to process real-time data using an FPGA / ASIC accelerator, calculate the rate of change of impedance and the rate of change of contact force, and provide data to the control algorithm and model calculation module. The nonlinear model calculation module is used to calculate the depth of ablation lesions using a neural network machine learning model, while dynamically updating model parameters, supporting online learning, and providing prediction results of the depth of ablation lesions. The control algorithm module is used to implement a PID controller, dynamically adjust the discharge parameters, optimize the combination of discharge parameters using a Bayesian optimization algorithm, and provide control signals to the electric field control module. The electric field control module is used to receive control signals, adjust the discharge parameters of the high-frequency pulse generator, adjust the impedance matching circuit in real time, optimize the electric field distribution, control the pulse discharge process, and ensure precise electric field control. The contact force adjustment module is used to dynamically adjust the contact state between the catheter and myocardial tissue based on the contact force sensor data, so as to ensure that the catheter and tissue are in full contact and optimize the electric field distribution. The safety protection module is used to monitor voltage, current and impedance parameters in real time, detect abnormal situations such as overvoltage, overcurrent and impedance sudden changes and trigger the protection mechanism, and provide safety alarms and handling suggestions.

3. The myocardial ablation catheter control system based on pulsed field ablation according to claim 2, characterized in that: The data acquisition module performs filtering calculations as follows: ； in: The filtered output at the current moment, This is the filtered output from the previous time step. The input data value at the current moment. Indicates the current input value Contribution to the current filtered output; The data standardization processing calculation is as follows: ； in: The original data, The standard deviation of the data; The mean of the data. These are the standardized data values.

4. The myocardial ablation catheter control system based on pulsed field ablation according to claim 3, characterized in that: The impedance change rate is calculated in the real-time data processing module as follows: ； Where: the rate of change of impedance is ； It is the impedance value at the current moment. It is the impedance value at the previous moment. It is a time interval, in seconds.

5. A myocardial ablation catheter control system based on pulsed field ablation according to claim 4, characterized in that: The rate of change of contact force in the real-time data processing module is calculated as follows: ； Among them: the rate of change of the contact force ； It is the current contact force value. It is the contact force value at the previous moment.

6. A myocardial ablation catheter control system based on pulsed field ablation according to claim 5, characterized in that: The neural network model in the nonlinear model calculation module is as follows: Assuming the neural network model is a multilayer perceptron (MLP), and the ablation lesion depth is... The prediction formula is: ； in: It is the input feature vector at the current moment. It is the weight matrix of the neural network. It is the bias vector of the neural network. It is a non-linear function of a neural network.

7. A myocardial ablation catheter control system based on pulsed field ablation according to claim 6, characterized in that: The online learning update calculation in the nonlinear model calculation module is as follows: During online learning, model parameters and Dynamically updated based on new data, the update is based on gradient descent: ； in: It is the mean squared error of the loss function. It's the learning rate. and These are the gradients of the loss function with respect to the weights and biases, respectively. For the updated model parameters ， For the updated model parameters 。 8. A myocardial ablation catheter control system based on pulsed field ablation according to claim 7, characterized in that: The depth of the ablation lesion is calculated in the nonlinear model calculation module as follows: Obtain input feature data from the real-time data processing module. ： ； Predict the depth of the ablation lesion using the current model parameters: ； Based on the new input data and the actual ablation depth Update model parameters: ； ； Predicted ablation depth Provided to the control algorithm or model calculation module.

9. A myocardial ablation catheter control system based on pulsed field ablation according to claim 8, characterized in that: The PID controller in the control algorithm module calculates as follows: ； in: For control signals; For proportional gain; For integral gain; This is the differential gain; This is the error signal, which is the difference between the set value and the actual value; The Bayesian optimization algorithm calculates as follows: ； in: This is a combination of discharge parameters; It is the mean vector; It is the covariance matrix; The Bayesian optimization algorithm also includes a posterior distribution, which is calculated as follows: ； in: These are the observed values; The observation matrix; Here is the noise covariance matrix; The optimal combination of discharge parameters is calculated as follows: ； in: The optimal combination of discharge parameters is obtained through Bayesian optimization; Indicates parameters It is a parameter vector; The control signal generation calculation is as follows: ； in: To provide control signals to the electric field control module; To be based on the optimal combination of discharge parameters The control signals for calculation.

10. A myocardial ablation catheter control system based on pulsed field ablation according to claim 9, characterized in that: The contact force adjustment calculation in the contact force adjustment module is as follows: ； in: For the desired fit; The contact force detected by the sensor in real time; To dynamically adjust the gain and optimize the contact state between the catheter and myocardial tissue; The dynamically adjusted gain is calculated as follows: ； in: This is the contact force adjustment coefficient; Adhesion to the target The optimal electric field distribution is calculated as follows: ； in: The optimized electric field distribution; Let be the electric field distribution function.

Images