Detecting atrial fibrillation and atrial fibrillation termination
A method is provided. The method is implemented by a detection engine embodied in processor executable code stored on a memory and executed by at least one processor. The method includes modeling a vector velocity field that measures and quantifies a velocity of electrocardiogram data signals that pass through a local activation time. The method further includes determining codes for each point in plane to provide a color code vector field image; detecting focal and rotor indications by using kernels to scan the color coded vector field image; and classifying the focal and rotor indications into perpetuators.
1 . A method for improving efficacy of a cardiac ablation procedure, the method comprising:
obtaining, from a mapping catheter within a heart of a patient, electrocardiogram data signals including intracardiac electrograms during the cardiac ablation procedure;
modeling a velocity vector field by fitting a polynomial surface T(x,y) to local activation time values at electrode locations, applying regularization, and computing, for each location, a two-dimensional velocity vector as a spatial derivative of T(x,y);
determining, for each point in a plane, a directional code selected from directional codes (left, right, up, down) according to respective signs of x- and y-components of the velocity vector field, and forming a color coded vector field image;
detecting focal and rotor indications by scanning the color coded vector field image with a circular kernel, wherein a rotor indication is detected when the directional codes within the circular kernel occur in a clockwise or counter-clockwise circular order around the circular kernel, and a focal indication is detected when the directional codes within the kernel point appear in order;
classifying the focal and the rotor indications into perpetuators based on the velocity vector field and ablation information collected during the cardiac ablation procedure, wherein an active perpetuator is a location at which ablation causes termination of atrial fibrillation or a cycle length prolongation exceeding a threshold, a passive perpetuator is a location at which ablation causes no visible change, and an unknown perpetuator is a location at which nearby ablation has not been performed; and
configuring, during the cardiac ablation procedure, one or more parameters of a surgical console that performs the cardiac ablation procedure, including at least one of power, duration, and contact force, based on the perpetuators.
2 . The method of claim 1 , further comprising:
detecting one or more segments of the local activation time values relative to a first activation time.
3 . The method of claim 1 , wherein the modeling includes calculating a direction of an electrical wave at each x, y point and utilizing a derivative of the polynomial surface to provide the velocity vector field.
4 . The method of claim 1 , wherein the perpetuators are classified by utilizing the velocity vector field as an input to a machine learning or artificial intelligence algorithm.
5 . The method of claim 4 , wherein the machine learning or the artificial intelligence algorithm comprises a deep convolutional neural network or a recurrent neural network to detect locations of gold standard perpetuators of the perpetuators.
6 . The method of claim 4 , wherein the machine learning or the artificial intelligence algorithm includes data identifying whether a prior cardiac ablation was successful on a respective patient.
7 . The method of claim 1 , further comprising: annotating atrial fibrillation perpetuators based on the vector velocity field and the ablation information.
8 . The method of claim 1 , wherein fitting the polynomial surface comprises minimizing a cost function that includes an L2 regularization term and performing gradient descent to obtain T(x,y), and wherein the velocity vector field is a gradient of T(x,y).
9 . The method of claim 1 , wherein the circular kernel has a radius of about 1 millimeter.
10 . The method of claim 1 , wherein classifying includes training a machine-learning model using labels derived from termination or cycle length prolongation as gold-standard outcomes.
11 . A system for improving efficacy of a cardiac ablation procedure, the system comprising:
a memory; and
one or more processors that are communicatively coupled to the memory, wherein the one or more processors are collectively configured to:
obtain, from a mapping catheter within a heart of a patient, electrocardiogram data signals comprising intracardiac electrograms during the cardiac ablation procedure;
model a velocity vector field by fitting a polynomial surface T(x,y) to local activation time values at electrode locations, applying regularization, and computing, for each location, a two-dimensional velocity vector as a spatial derivative of T(x,y);
determine, for each point in a plane, a directional code selected from four codes (left, right, up, down) to provide a color coded vector field image;
detect focal and rotor indications by using one or more circular kernels scan the color coded vector field image, detecting a rotor indication only upon a clockwise or counter-clockwise circular order of the directional codes within a kernel;
classify the focal and the rotor indications into perpetuators based on the velocity vector field and ablation information collected during the cardiac ablation procedure, including labeling active perpetuators by termination or cycle length prolongation after ablation; and
automatically configure, during the cardiac ablation procedure, at least one of ablation power, duration, and contact force of a surgical console based on the perpetuators.
12 . The system of claim 11 , wherein the one or more processors are further collectively configured to:
detect one or more segments of the local activation time values relative to a first activation time.
13 . The system of claim 11 , wherein the velocity field is modeled by calculating a direction of an electrical wave at each x, y point and utilizing a derivative of the polynomial surface to provide the velocity vector field.
14 . The system of claim 11 , wherein the perpetuators are classified by utilizing the velocity vector field as an input to a machine learning or artificial intelligence algorithm.
15 . The system of claim 14 , wherein the machine learning or the artificial intelligence algorithm comprises a deep convolutional neural network or a recurrent neural network to detect locations of gold standard perpetuators of the perpetuators.
16 . The system of claim 14 , wherein the machine learning or the artificial intelligence algorithm includes data identifying whether a prior cardiac ablation was successful on a respective patient.
17 . The system of claim 11 , wherein the one or more processors are further collectively configured to:
automatically annotate atrial fibrillation perpetuators based on the vector velocity field and the ablation information.
18 . A non-transitory computer readable medium storing instructions for improving efficacy of a cardiac ablation procedure, the instructions when executed by a surgical console that performs the cardiac ablation procedure, cause the surgical console to perform a method comprising:
obtaining, from a mapping catheter within a heart of a patient, electrocardiogram data signals comprising intracardiac electrograms during the cardiac ablation procedure;
modeling a velocity vector by fitting a polynomial surface to local activation time values with regularization and computing a two-dimensional velocity vector as a spatial derivative;
determining, for each point in a plane to provide, a four-state directional code and forming a color coded vector field image;
detecting focal and rotor indications by using circular kernels to scan the color coded vector field image, detecting a rotor indication only upon a clockwise or counter-clockwise circular order of the directional codes within a kernel;
classifying the focal and the rotor indications into perpetuators based on the velocity vector field and ablation information collected during the cardiac ablation procedure, including labeling active perpetuators by termination or cycle length prolongation after ablation; and
configuring, during the cardiac ablation procedure, one or more parameters of the surgical console, including at least one of power, duration, and contact force, based on the perpetuators.