atrial fibrillation

Abstract

To provide a system and method for implementing a machine-learning / artificial-intelligence algorithm for detecting atrial fibrillation and atrial fibrillation termination.SOLUTION: A method is provided. The method is implemented by a detection engine embodied in a 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 a plane to provide a color-coded vector field image; detecting focal activation and rotor indications by using kernels to scan the color-coded vector field image; and classifying the focal activation and rotor indications into perpetuators.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] (Priority Claim) This application claims priority to U.S. Provisional Patent Application No. 63 / 039,291, entitled "ATRIAL FIBRILLATION," which is incorporated by reference for all purposes as if set forth in its entirety herein.

[0002] FIELD OF THE INVENTION The present invention relates to artificial intelligence and machine learning related to atrial fibrillation, and more particularly to systems and methods for implementing machine learning / artificial intelligence algorithms for detecting atrial fibrillation and atrial fibrillation termination. [Background technology]

[0003] Atrial fibrillation ("aFib") is a quivering or irregular heartbeat (arrhythmia) that can lead to blood clots, stroke, heart failure, and other heart-related complications. aFib is the most common arrhythmia diagnosed in clinical practice. Estimates of the prevalence of aFib in the United States range from approximately 2.7 million to 6.1 million, and are expected to rise to 12.1 million by 2030. According to a 2013 study, global estimates place the number of individuals with aFib at approximately 33.5 million in 2010, representing approximately 0.5 percent of the world's population.

[0004] Currently, physicians cannot determine whether and when a particular aFib case has been successfully terminated from a set of electrocardiogram (ECG) signals. For example, conventional mechanisms consider phase mapping (i.e., calculating interelectrode delays and finding phase singularities) and vector analysis without using velocity vector fields, machine learning, deep learning, and / or other detailed algorithms. At best, conventional mechanisms can use velocity characteristics of aFib sources determined by flow mapping of electrograms before and after catheter ablation. Even so, these conventional mechanisms cannot distinguish between active and inactive focal sources. Similarly, conventional mechanisms are far from ideal in determining aFib termination. Summary of the Invention [Means for solving the problem]

[0005] According to an exemplary embodiment, a method is provided. The method is implemented by a detection engine embodied as 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 the velocity of an electrocardiogram data signal through a local activation time. The method further includes determining one or more codes for each point in a plane to provide a color-coded vector field image, detecting focal activation and rotor indices by scanning the color-coded vector field image using one or more kernels, and classifying the focal activation and rotor indices into perpetuators.

[0006] According to one or more embodiments, the exemplary method embodiments described above may be implemented as an apparatus, a system, and / or a computer program product. [Brief explanation of the drawings]

[0007] A more detailed understanding may be had from the following description, given by way of example in conjunction with the accompanying drawings, in which like reference numerals refer to like elements in the various drawings, and in which: [Figure 1] 1 shows a diagram of an exemplary system capable of implementing one or more features of the subject matter of this disclosure, according to one or more embodiments. [Figure 2] FIG. 1 illustrates a block diagram of an exemplary system for anatomically accurate reconstruction of the atria of the heart, in accordance with one or more embodiments. [Figure 3] 1 illustrates a method according to one or more embodiments. [Figure 4] 1 depicts a pictorial representation of an artificial intelligence system in accordance with one or more illustrative embodiments. [Figure 5] 1 illustrates a block diagram of an example neural network and a method implemented within the neural network, according to one or more embodiments. [Figure 6] 1 illustrates a method according to one or more embodiments. [Figure 7] 1 illustrates a graph according to one or more embodiments. [Figure 8] 1 illustrates a graph of a surface (x,y) in accordance with one or more embodiments. [Figure 9] 9 illustrates a graph of a velocity vector field for the surface (x,y) of FIG. 8 in accordance with one or more embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0008] Disclosed herein are methods and systems for machine learning and / or artificial intelligence implemented by a detection engine. More particularly, the present invention relates to a detection engine that includes machine learning / artificial intelligence algorithms for detecting aFib and aFib termination.

[0009] One or more advantages, technical effects, and / or effects of the detection engine may include solving activity and inactivity questions (e.g., distinguishing between focal sources and rotors) using a feedback loop based on actual ablation of focal activation. In this regard, the detection engine may model a vector velocity field while distinguishing between focal sources and rotors (e.g., active and inactive). That is, the detection engine may define where to calculate local activation time (LAT), calculate the derivative of the LAT to obtain velocity, identify focal sources (e.g., within a 1 mm box), classify focal sources as active or inactive (e.g., within a 1 mm box), provide retrospective analysis (e.g., review past cases and determine where ablation occurred), and provide prospective analysis (e.g., determine where ablation should occur).

[0010] For ease of explanation, the detection engine is described herein with respect to determining and treating aFib in the heart, however, any anatomical structure, body part, organ, or portion thereof may be targeted for mapping by the detection engine described herein. Furthermore, the detection engine and / or machine learning / artificial intelligence algorithms are processor-executable code or software that may be subject to processing operations by, and not necessarily routed to, the processing hardware of, a medical device device.

[0011] According to one or more embodiments, a method is provided that is implemented by a detection engine embodied as processor-executable code stored on a memory and executed by at least one processor, the method including: modeling, by the detection engine, a vector velocity field that measures and quantifies the velocity of an electrocardiogram data signal through a local activation time; determining, by the detection engine, one or more codes for each point in a plane to provide a color-coded vector field image; detecting, by the detection engine, focal activation and rotor indices by scanning the color-coded vector field image using one or more kernels; and classifying, by the detection engine, the focal activation and rotor indices into persistence factors.

[0012] According to one or more of the embodiments or method embodiments herein, electrocardiogram data signals may be detected by a catheter within the anatomy in communication with a detection engine.

[0013] According to one or more embodiments or any of the method embodiments herein, the detection engine may detect one or more segments of local excitation times relative to the first excitation time.

[0014] According to one or more of the embodiments or method embodiments herein, the detection engine can calculate the direction of the radio waves at each x,y point and model the vector velocity field by using derivatives of a polynomial surface to give the velocity vector field.

[0015] According to any of one or more embodiments or method embodiments herein, the detection engine can classify the persistence factor by using the velocity vector field as an input to a machine learning or artificial intelligence algorithm.

[0016] According to any of the one or more embodiments or method embodiments herein, the machine learning or artificial intelligence algorithm can include a deep convolutional neural network or a recurrent neural network to detect the location of the gold standard persistence factor among the persistence factors.

[0017] According to any of one or more embodiments or method embodiments herein, a machine learning or artificial intelligence algorithm can determine whether the ablation outcome for a particular case related to an electrocardiogram data signal was successful.

[0018] According to any of one or more embodiments or method embodiments herein, focal excitation and rotor indices can be detected when all directions are in order within one or more kernels.

[0019] According to any of one or more embodiments or method embodiments herein, the detection engine can automatically identify and annotate atrial fibrillation persistence factors based on vector velocity and ablation information when marking persistence factors.

[0020] According to any of one or more embodiments or method embodiments herein, annotations of the region of interest can be used by the detection engine with respect to active, inactive, and unknown categories to indicate at least whether the region of interest has been ablated or whether atrial fibrillation has been terminated.

[0021] According to one or more embodiments, the system includes a memory storing processor-executable code for a detection engine. The system also includes at least one processor that executes the processor-executable code to cause the system to: model, with the detection engine, a vector velocity field that measures and quantifies the velocity of the electrocardiogram data signal through a local activation time; determine, with the detection engine, one or more codes for each point in the plane to provide a color-coded vector field image; detect, with the detection engine, focal activation and rotor indices by scanning the color-coded vector field image using one or more kernels; and classify, with the detection engine, the focal activation and rotor indices into persistence factors.

[0022] According to one or more embodiments or any of the system embodiments herein, electrocardiogram data signals may be detected by a catheter within the anatomy in communication with a detection engine.

[0023] According to any of one or more embodiments or system embodiments herein, the detection engine may detect one or more segments of the local excitation time relative to the first excitation time.

[0024] According to one or more of the embodiments or system embodiments herein, the detection engine can calculate the direction of the radio waves at each x,y point and model the vector velocity field by using derivatives of a polynomial surface to give the velocity vector field.

[0025] According to one or more embodiments or any of the system embodiments herein, the detection engine can classify the persistence factor by using the velocity vector field as an input to a machine learning or artificial intelligence algorithm.

[0026] According to any of the one or more embodiments or system embodiments herein, the machine learning or artificial intelligence algorithm can include a deep convolutional neural network or a recurrent neural network to detect the location of the gold standard persistence factor among the persistence factors.

[0027] According to any of one or more embodiments or system embodiments herein, a machine learning or artificial intelligence algorithm can determine whether the ablation outcome for a particular case related to the electrocardiogram data signal was successful.

[0028] According to any of one or more embodiments or system embodiments herein, focal excitation and rotor indices can be detected when all directions are in order within one or more kernels.

[0029] According to one or more embodiments or any of the system embodiments herein, the detection engine can automatically identify and annotate atrial fibrillation persistence factors based on vector velocity and ablation information when marking persistence factors.

[0030] According to one or more embodiments or any of the system embodiments herein, annotations of the region of interest can be used by the detection engine with respect to active, inactive, and unknown categories to indicate at least whether the region of interest has been ablated or whether atrial fibrillation has been terminated.

[0031] 1 is a diagram of an exemplary system (e.g., a medical device) shown as system 100 in which one or more features of the subject matter herein may be implemented in accordance with one or more embodiments. All or a portion of system 100 may be used to collect information (e.g., biometric data and / or training data sets) and / or implement detection engine 101 (e.g., machine learning and / or artificial intelligence algorithms) described herein. Detection engine 101 may be defined as a deep learning optimization for detecting persistence factors of aFibs to be ablated to treat persistent aFib patients and classifying velocity vector field images and raw data into persistence factors.

[0032] The illustrated system 100 includes a probe 105 with a catheter 110 (including at least one electrode 111), a shaft 112, a sheath 113, and a manipulator 114. The illustrated system 100 also includes a physician 115 (or medical professional or clinician), a heart 120, a patient 125, and a bed 130 (or table). Note that insets 140 and 150 show the heart 120 and catheter 110 in greater detail. The system 100 also includes a console 160 (including one or more processors 161 and memory 162) and a display 165, as shown. Further, note that each element and / or item of the system 100 represents one or more of that element and / or item. The example system 100 illustrated in FIG. 1 may be modified to implement embodiments disclosed herein. The disclosed embodiments may be similarly applied using other system components and configurations. Additionally, system 100 may include additional components, such as elements for sensing electrical activity, wired or wireless connectors, processing and display devices, etc.

[0033] System 100 can be used to detect, diagnose, and / or treat cardiac conditions (e.g., using detection engine 101). Cardiac conditions, such as cardiac arrhythmias, remain common and dangerous medical disorders, particularly in the elderly population. For example, system 100 can be part of a surgical system (e.g., the CARTO® system sold by Biosense Webster) configured to acquire biometric data (e.g., anatomical and electrical measurements of a patient's organs, such as heart 120) and perform cardiac ablation procedures. More specifically, treating cardiac disorders, such as cardiac arrhythmias, often requires detailed mapping of cardiac tissue, chambers, veins, arteries, and / or electrical pathways. For example, as a prerequisite for successful catheter ablation (as described herein), the source of the cardiac arrhythmia may be accurately localized in a chamber of heart 120. Such localization can be performed by electrophysiological testing, in which spatially resolved electrical potentials are detected by a mapping catheter (e.g., catheter 110) introduced into a chamber of heart 120. This electrophysiological test, also known as electroanatomical mapping, therefore provides 3D mapping data that can be displayed on a monitor. Often, the mapping and treatment functions (e.g., ablation) are provided by a single catheter or a group of catheters, whereby the mapping catheter also simultaneously operates as a treatment (e.g., ablation) catheter. In this case, the detection engine 101 can be stored and executed directly by the catheter 110.

[0034] It should be noted that in a patient (e.g., patient 125) having normal sinus rhythm (NSR), the heart (e.g., heart 120), including the atria, ventricles, and excitable conduction tissue, is electrically excited to beat in a synchronized, patterned manner, which can be detected, for example, as intracardiac electrocardiogram (IC ECG) data.

[0035] The heart 120 typically consists of four chambers: two upper chambers (atria) and two lower chambers (ventricles). The coronary sinus (CS) is a collection of converging veins that form a large vessel that collects blood from the myocardium and sends less-oxygenated blood to the right atrium. The rhythm of the heart 120 is typically controlled by the sinus node (not shown), located in the right atrium. The sinus node typically generates an electrical impulse that initiates each heartbeat and acts as a natural pacemaker. The electrical impulse travels from the sinus node through the atria, causing the atrial muscle to contract and pump blood into the ventricles. The electrical impulse then reaches a group of cells called the atrioventricular node (AV node) (not shown). The AV node is typically the only pathway for signals to travel from the atria to the ventricles. The AV node slows down the electrical impulse before sending it to the ventricles. This delay, even if brief, allows the ventricles to fill with blood. When an electrical impulse reaches the ventricular muscles, they contract, pumping blood to the lungs or the rest of the body. In a healthy heart 120, this process typically proceeds smoothly, resulting in a normal resting heart rate of 60 to 100 beats per minute. In a heart 120 with one of the disease conditions identified above, faulty electrical connections or abnormal areas of electrical activity within the heart induce and perpetuate an abnormal rhythm. When this occurs, the heart rate rapidly accelerates, not allowing the heart 120 enough time to fill before contracting again. These ineffective contractions of the heart 120 can cause lightheadedness or dizziness, as the brain may not receive enough blood and oxygen.

[0036] In patients (e.g., patient 125) with cardiac arrhythmias (e.g., atrial fibrillation or aFib), abnormal regions of cardiac tissue do not follow the synchronous beating cycle associated with normal conductive tissue, in contrast to patients with NSR. Instead, the abnormal regions of cardiac tissue experience aberrant conduction to adjacent tissue, disrupting the cardiac cycle and resulting in asynchronous cardiac rhythms. Note that this asynchronous cardiac rhythm can also be detected in IC ECG data. Such abnormal conduction has previously been known to occur in various regions of the heart 120, such as in the region of the sinoatrial (SA) node along the atrioventricular (AV) node's conduction pathway, or in the myocardial tissue forming the walls of the ventricles and atria. Other conditions, such as atrial flutter, exist in which abnormal conductive tissue patterns lead to reentry pathways, causing the heart chambers to beat in a regular pattern that can be multiples of sinus rhythm.

[0037] To assist system 100 in detecting, diagnosing, and / or treating a cardiac condition, physician 115 can guide probe 105 into heart 120 of patient 125 reclining on bed 130. For example, physician 115 can insert shaft 112 through sheath 113 while manipulating the distal end of shaft 112 using manipulator 114 near the proximal end of catheter 110 and / or deflection from sheath 113. As shown in inset 140, catheter 110 can be attached to the distal end of shaft 112. Catheter 110 can be inserted through sheath 113 in a collapsed state and then expanded within heart 120.

[0038] In general, electrical activity at a point within the heart 120 can typically be measured by advancing a catheter 110 (e.g., at least one electrode 111) containing an electrical sensor at or near its distal tip to the point within the heart 120, contacting tissue with the sensor, and collecting data at the point. One difficulty with mapping a heart chamber using a catheter containing only a single distal tip electrode is the long time required to collect data for each point across the necessary number of points required for a detailed map of the entire heart chamber. Therefore, multi-electrode catheters (e.g., catheter 110) have been developed to simultaneously measure electrical activity at multiple points within a heart chamber.

[0039] The catheter 110, which may include at least one electrode 111 and a catheter needle coupled on its body, may be configured to obtain biometric data, such as electrical signals, of a body organ (e.g., the heart 120) and / or ablate a tissue region thereof (e.g., a chamber of the heart 120). It should be noted that the electrode 111 represents any similar element, such as a tracking coil, piezoelectric transducer, electrode, or combination of elements configured to ablate a tissue region or obtain biometric data. According to one or more embodiments, the catheter 110 may include one or more position sensors used to determine trajectory information. This trajectory information may be used to infer motion characteristics, such as tissue contractility.

[0040] The biometric data (e.g., patient biometrics, patient data, or patient biometric data) may include one or more of local activation time (LAT), electrical activity, topology, bipolar mapping, baseline activity, ventricular activity, dominant frequency, impedance, etc. LAT may be a time point of threshold activity corresponding to local activation calculated based on a normalized initial onset. Electrical activity may be any applicable electrical signal that can be measured based on one or more thresholds and may be detected and / or enhanced based on signal-to-noise ratio and / or other filters. Topology may correspond to the physical structure of a body part or portion of a body part, and may correspond to changes in the physical structure for different portions of the body part or for different body parts. The dominant frequency may be a frequency or range of frequencies commonly found in a portion of a body part and may differ in different portions of the same body part. For example, the dominant frequency of the PVs of a heart may differ from the dominant frequency of the right atrium of the same heart. Impedance may be a resistance measurement in a particular region of a body part.

[0041] Examples of biometric data include, but are not limited to, patient identification data, IC ECG data, bipolar intracardiac reference signals, anatomical and electrical measurements, trajectory information, body surface (BS) ECG data, historical data, brain biometrics, blood pressure data, ultrasound signals, radio signals, audio signals, two-dimensional or three-dimensional image data, blood glucose data, and temperature data. Biometric measurements may generally be used to monitor, diagnose, and treat any number of various diseases, such as cardiovascular diseases (e.g., arrhythmias, cardiomyopathies, and coronary artery disease) and autoimmune diseases (e.g., type I and type II diabetes). Note that BS ECG data may include data and signals collected from electrodes on the patient's surface, IC ECG data may include data and signals collected from electrodes inside the patient's body, and ablation data may include data and signals collected from ablated tissue. Furthermore, BS ECG data, IC ECG data, and ablation data, along with catheter electrode position data, may be derived from one or more treatment records.

[0042] For example, catheter 110 may use electrodes 111 to perform intravascular ultrasound and / or MRI catheterization to image (e.g., acquire and process biometric data for) heart 120. Inset 150 shows a close-up view of catheter 110 inside a chamber of heart 120. While catheter 110 is shown as a point catheter, it will be understood that any shape that includes one or more electrodes 111 may be used to implement the exemplary embodiments disclosed herein.

[0043] Examples of the catheter 110 include, but are not limited to, a straight catheter with multiple electrodes, a balloon catheter including electrodes distributed on multiple spines forming a balloon, a lasso or loop catheter with multiple electrodes, a force-sensing catheter, or any other applicable shape or type. The straight catheter may be fully or partially elastic so that it can twist, bend, and / or otherwise change its shape based on received signals and / or the action of an external force (e.g., cardiac tissue) on the straight catheter. The balloon catheter may be designed to hold its electrodes in intimate contact with the endocardial surface when deployed within a patient's body. As an example, the balloon catheter may be inserted into a lumen such as the PV. The balloon catheter may be inserted into the PV in a deflated state so that the balloon catheter does not occupy its maximum volume while inserted into the PV. The balloon catheter may be expanded while inside the PV so that the electrodes on the balloon catheter contact the entire circular portion of the PV. Such contact with the entire circular portion of the PV or any other lumen enables efficient imaging and / or ablation.

[0044] According to another example, body patches and / or body surface electrodes may be positioned on or adjacent to the body of the patient 125. A catheter 110 having one or more electrodes 111 may be positioned within the body (e.g., within the heart 120), and the position of the catheter 110 may be determined by the system 100 based on signals transmitted and received between the one or more electrodes 111 of the catheter 110 and the body patches and / or body surface electrodes. Additionally, the electrodes 111 may sense biometric data from within the body of the patient 125, such as within the heart 120 (e.g., the electrodes 111 sense tissue electrical potentials in real time). The biometric data may be correlated with the determined position of the catheter 110, thereby allowing a rendering of the patient's body part (e.g., the heart 120) to be displayed, showing the biometric data superimposed on the shape of the body part.

[0045] The probe 105 and other items of the system 100 can be connected to a console 160. The console 160 can include any computing device that employs machine learning and / or artificial intelligence algorithms (represented as the detection engine 101). According to an exemplary embodiment, the console 160 includes one or more processors 161 (any computing hardware) and memory 162 (any non-transitory tangible medium), where the one or more processors 161 execute computer instructions for the detection engine 101 and the memory 162 stores these instructions for execution by the one or more processors 161. For example, the console 160 can be configured to receive and process biometric data to determine whether a particular tissue region conducts electricity.

[0046] In some embodiments, the console 160 can be further programmed by the detection engine 101 (in software) to perform the functions of: modeling a vector velocity field that measures and quantifies the velocity of electrocardiogram data signals through local activation times; determining one or more codes for each point in a plane to provide a color-coded vector field image; detecting focal activation and rotor indices by scanning the color-coded vector field image using one or more kernels (e.g., space in a map); and classifying the focal activation and rotor indices into persistence factors. For example, the detection engine 101 can include deep learning optimization (described herein in connection with FIGS. 3 and 6 ) that receives biometric data acquired by the catheter 110 as it is maneuvered within the anatomical structure. After the mapping is generated, the detection engine 101 can receive input representing user modifications of the mapping, such as via an existing user interface and / or a specialized user interface of the detection engine 101. Generally, the detection engine 101 can provide one or more user interfaces, such as in place of an operating system or other application and / or directly as needed. User interfaces may include, but are not limited to, an internet browser, a graphic user interface (GUI), a windowing interface, and / or other visual interfaces for applications, operating systems, file folders, etc. According to one or more embodiments, the detection engine 101 may be external to the console 160, for example, located within the catheter 110, an external device, a mobile device, a cloud-based device, or may be a stand-alone processor. In this regard, the detection engine 101 may be transferable / downloadable in electronic form over a network.

[0047] In one example, console 160 may be any computing device described herein, including software (e.g., detection engine 101) and / or hardware (e.g., processor 161 and memory 162), such as a general-purpose computer with appropriate front-end and interface circuitry for sending and receiving signals to and from probe 105 and for controlling other components of system 100. For example, the front-end and interface circuitry may include an input / output (I / O) communication interface that allows console 160 to receive signals from and / or transmit signals to at least one electrode 111. Console 160 may include real-time noise reduction circuitry, typically configured as a field-programmable gate array (FPGA) followed by an analog-to-digital (A / D) ECG or electrocardiogram / electromyogram (EMG) signal conversion integrated circuit. Console 160 may communicate signals from the A / D ECG or EMG circuitry to a separate processor and / or may be programmed to perform one or more functions disclosed herein.

[0048] A display 165, which may be any electronic device for visually presenting biometric data, is connected to the console 160. According to an exemplary embodiment, during a procedure, the console 160 may facilitate the presentation of a rendering of the body part to the physician 115 on the display 165 and store data representing the rendering of the body part in the memory 162. For example, a map indicative of motion characteristics may be rendered / constructed based on trajectory information sampled at a sufficient number of points within the heart 120. As an example, the display 165 may include a touch screen, which may be configured to receive input from the medical professional 115 in addition to presenting the rendering of the body part.

[0049] In some exemplary embodiments, physician 115 can use one or more input devices, such as a touchpad, mouse, keyboard, gesture recognizer, etc., to manipulate renderings of elements of system 100 and / or body parts. For example, the input devices can be used to change the position of catheter 110 so that renderings are updated. Note that display 165 can be located at the same location or at a remote location, such as another hospital or another healthcare provider network.

[0050] According to one or more embodiments, the system 100 can also obtain biometric data using ultrasound, computed tomography (CT), MRI, or other medical imaging techniques utilizing the catheter 110 or other medical equipment. For example, the system 100 can obtain ECG data and / or anatomical and electrical measurements (e.g., biometric data) of the heart 120 using one or more catheters 110 or other sensors. More specifically, the console 160 can be connected by cable to BS electrodes, including adhesive skin patches, attached to the patient 125. The BS electrodes can acquire / generate biometric data in the form of BS ECG data. For example, the processor 161 can determine position coordinates of the catheter 110 within a body portion (e.g., the heart 120) of the patient 125. The position coordinates can be based on impedance or electromagnetic fields measured between body surface electrodes and electrodes 111 of the catheter 110 or other electromagnetic components. Additionally or alternatively, location pads that generate magnetic fields used for navigation can be placed on the surface of the bed 130 or can be separate from the bed 130. The biometric data may be transmitted to the console 160 and stored in memory 162. Alternatively or additionally, the biometric data may be transmitted to a server, which may be local or remote, using a network as described further herein.

[0051] According to one or more exemplary embodiments, catheter 110 can be configured to ablate a tissue region of a chamber of heart 120. Inset 150 shows a close-up view of catheter 110 inside a chamber of heart 120. For example, an ablation electrode, such as at least one electrode 111, can be configured to apply energy to a tissue region of an internal organ (e.g., heart 120). The energy can be thermal energy and can cause damage to the tissue region starting from the surface of the tissue region and extending through the thickness of the tissue region. Biometric data related to the ablation procedure (e.g., ablated tissue, ablation location, etc.) can be considered ablation data.

[0052] According to one example, with respect to acquiring biometric data, a multi-electrode catheter (e.g., catheter 110) can be advanced into a chamber of heart 120. An anterior-posterior (AP) and lateral fluoroscopic photograph can be acquired to determine the position and orientation of each of the electrodes. An ECG can be recorded from each of the electrodes 111 in contact with the cardiac surface relative to a time reference, such as the occurrence of P waves in sinus rhythm from a BS ECG and / or a signal from an electrode 111 of catheter 110 positioned within the coronary sinus. Systems further disclosed herein can distinguish between electrodes that record electrical activity and electrodes that do not record electrical activity due to their lack of proximity to the endocardial wall. After an initial ECG is recorded, the catheter can be repositioned, and fluoroscopic photographs and an ECG can be recorded again. An electrical map (e.g., via cardiac mapping) can then be constructed from an iteration of the above process.

[0053] Cardiac mapping can be performed using one or more techniques. Generally, mapping of cardiac regions, such as the cardiac regions, tissues, veins, arteries, and / or electrical pathways of the heart 120, can lead to the identification of problem areas, such as scar tissue, arrhythmia sources (e.g., electrical rotors), healthy regions, etc. Cardiac regions can be mapped such that a visual rendering of the mapped cardiac region is provided using a display, as further disclosed herein. Furthermore, cardiac mapping (which is an example of cardiac imaging) may include mapping based on one or more modalities, such as, but not limited to, LAT, regional activation rate, electrical activity, topology, bipolar mapping, dominant frequency, or impedance. Data (e.g., biometric data) corresponding to multiple modalities can be captured using a catheter (e.g., catheter 110) inserted into the patient's body and can be provided for rendering simultaneously or at different times based on corresponding settings and / or physician 115 preferences.

[0054] As an example of the first technique, cardiac mapping can be performed by sensing electrical properties of cardiac tissue, such as LAT, as a function of precise location within the heart 120. Corresponding data (e.g., biometric data) can be acquired by one or more catheters (e.g., catheter 110) having electrical and position sensors (e.g., electrodes 111) at their distal tips that are advanced into the heart 120. By way of example, location and electrical activity can typically be initially measured at about 10 to about 20 points on the inner surface of the heart 120. These data points are generally sufficient to generate a preliminary reconstruction or map of the cardiac surface of satisfactory quality. This preliminary map is then combined with data measured at additional points to generate a more comprehensive map of the cardiac electrical activity. In clinical practice, it is not uncommon to collect data from 100 or more sites (e.g., several thousand) to generate a detailed, comprehensive map of the cardiac chamber's electrical activity. The detailed map can then serve as the basis for making decisions about therapeutic interventions, such as tissue ablation as described herein, to alter the propagation of the cardiac electrical activity and restore normal cardiac rhythm.

[0055] Furthermore, cardiac mapping can be generated based on the detection of intracardiac potential fields (e.g., IC ECG data and / or bipolar intracardiac reference signals, which are examples). Non-contact techniques can be implemented to simultaneously acquire large amounts of cardiac electrical information. For example, a catheter type having a distal end portion can include a series of sensor electrodes distributed over its surface and connected to insulated conductors for connection to signal sensing and processing means. The size and shape of the end portion can be such that the electrodes are spaced a large distance from the wall of the cardiac chamber. The intracardiac potential fields can be detected during a single heartbeat. According to one example, the sensor electrodes can be distributed on a series of circumferentially spaced planes. These planes can be perpendicular to the longitudinal axis of the catheter end portion. At least two additional electrodes can be provided adjacent each end of the longitudinal axis of the end portion. As a more specific example, the catheter can include four circumferences with eight electrodes equiangularly spaced on each circumference. Thus, in this particular implementation, the catheter can include at least 34 electrodes (32 circumferential electrodes and two distal electrodes). As another more specific example, the catheter may include other multi-spline catheters such as a five soft flexible branch, eight radial splines, or a turner type with parallel splines (e.g., any of which may have a total of 42 electrodes).

[0056] As an example of electrical or cardiac mapping, electrophysiological cardiac mapping systems and techniques based on non-contact and non-expandable multi-electrode catheters (e.g., catheter 110) can be implemented. An ECG can be obtained using one or more catheters 110 with multiple electrodes (e.g., 42 to 122 electrodes, etc.). According to this implementation, knowledge of the relative geometry of the probe and endocardium can be obtained by an independent imaging modality, such as transesophageal echocardiography. After the independent imaging, cardiac surface potentials can be measured using non-contact electrodes, and a map can be constructed from these surface potentials (e.g., possibly using a bipolar intracardiac reference signal). This technique can include (after the independent imaging step): (a) measuring potentials using multiple electrodes disposed on a probe placed within the heart 120; (b) determining the geometric relationship between the probe surface and the endocardial surface and / or other fiducials; (c) generating a matrix of coefficients representing the geometric relationship between the probe surface and the endocardial surface; and (d) determining the endocardial potentials based on the electrode potentials and the matrix of coefficients.

[0057] As another example of electrical or cardiac mapping, techniques and devices can be implemented for mapping the electrical potential distribution of a heart chamber. An intracardiac multi-electrode mapping catheter assembly can be inserted into heart 120. The mapping catheter (e.g., catheter 110) assembly can include a multi-electrode array or companion reference catheter with one or more integral reference electrodes (e.g., one or more electrodes 111).

[0058] According to one or more exemplary embodiments, the electrodes may be deployed in a generally spherical array, which may be spatially referenced to points on the endocardial surface by a reference electrode or by a reference catheter that is brought into contact with the endocardial surface. A preferred electrode array catheter may have a large number of individual electrode sites (e.g., at least 24). Additionally, this example technique may be implemented by knowing the location of each of the electrode sites on the array and by knowing the cardiac geometry. These locations are preferably determined by impedance plethysmography.

[0059] From an electrical or cardiac mapping perspective, and according to another example, the catheter 110 can be a cardiac mapping catheter assembly that can include an electrode array defining multiple electrode sites. The cardiac mapping catheter assembly also includes a lumen for receiving a reference catheter having a distal tip electrode assembly that can be used to probe the heart wall. The cardiac mapping catheter assembly can include a braid of insulated wires (e.g., having 24 to 64 wires within the braid), each of which can be used to form an electrode site. The cardiac mapping catheter assembly can be readily positioned within the heart 120 to be used to acquire electrical activity information from a first set of non-contact electrode sites and / or a second set of contact electrode sites.

[0060] According to another example, a catheter 110 capable of mapping electrophysiological activity within the heart may include a distal tip adapted to deliver stimulation pulses for pacing the heart or an ablation electrode for ablating tissue in contact with the tip, and may further include at least one pair of orthogonal electrodes for generating a differential signal indicative of local cardiac electrical activity in the vicinity of the orthogonal electrodes.

[0061] As described herein, system 100 can be used to detect, diagnose, and / or treat cardiac conditions. In an exemplary operation, system 100 can implement a process for measuring electrophysiological data within a heart chamber. This method can include, in part, positioning a set of active and passive electrodes within heart 120, applying current to the active electrodes to thereby generate an electric field within the heart chamber, and measuring the electric field at the passive electrode sites. The passive electrodes are included in an array disposed on an inflatable balloon of a balloon catheter. In a preferred embodiment, the array is said to have 60-64 electrodes.

[0062] As another exemplary operation, cardiac mapping may be performed by system 100 using one or more ultrasound transducers. The ultrasound transducers may be inserted into a patient's heart 120 and may collect multiple ultrasound slices (e.g., two-dimensional or three-dimensional slices) at various positions and orientations within heart 120. The position and orientation of a particular ultrasound transducer may be known, and the collected ultrasound slices may be stored for later display. One or more ultrasound slices corresponding to the position of probe 105 (e.g., a treatment catheter shown as catheter 110) may be displayed, and probe 105 may be overlaid on one or more ultrasound slices.

[0063] Considering system 100, it can be seen that cardiac arrhythmias, including atrial arrhythmias, can be multiwavelet-reentrant, characterized by multiple asynchronous loops of electrical impulses scattered around the atria, often self-propagating (e.g., another example of IC-ECG data). Alternatively or in addition to multiwavelet-reentrant, cardiac arrhythmias can also have focal excitation sources, such as when isolated regions of atrial tissue are autonomously excited in a rapid and repetitive manner (e.g., another example of IC-ECG data). Ventricular tachycardia (V-tach or VT) is a tachycardia or fast cardiac rhythm that occurs in one of the ventricles. It is a potentially fatal arrhythmia because it can lead to ventricular fibrillation and sudden death.

[0064] For example, aFib occurs when the normal electrical impulses generated by the sinoatrial node (e.g., another example of IC-ECG data) are overwhelmed by chaotic electrical impulses originating in the atrial veins and PVs (e.g., signal interference), resulting in irregular impulses being conducted to the ventricles. This results in an irregular heartbeat that may persist for minutes to weeks, or even years. In many cases, aFib is a chronic condition that often carries a small increase in the risk of death from stroke. The treatment approach for aFib is medication to reduce the heart rate or restore normal cardiac rhythm. Furthermore, patients with aFib are often given anticoagulants to protect against stroke. The use of such anticoagulants carries its own risks of internal bleeding. In some patients, medication is insufficient, and their aFib is deemed drug-refractory, i.e., untreatable by standard pharmacological intervention. Synchronized electrical cardioversion can also be used to convert aFib to a normal cardiac rhythm. Alternatively, patients with aFib may be treated with catheter ablation.

[0065] Treatment using catheter ablation may include mapping the electrical properties of cardiac tissue, particularly the endocardium and cardiac volumes, and selectively ablating cardiac tissue through the application of energy. Electrical or cardiac mapping (e.g., performed by any of the electrophysiological cardiac mapping systems and techniques described herein) involves creating an electrical potential map (e.g., a voltage map) of wave propagation along cardiac tissue or a map of arrival times (e.g., a LAT map) to points located within various tissues. Electrical or cardiac mapping (e.g., a cardiac map) can be used to detect localized cardiac tissue dysfunction. Ablation, such as cardiac mapping-based ablation, can stop or alter the propagation of unwanted electrical signals from one portion of the heart 120 to another.

[0066] Ablation techniques disrupt unwanted electrical pathways by creating non-conductive lesions. Various energy delivery modalities have been previously disclosed for creating lesions, including the use of microwave, laser, and more generally, radiofrequency energy to create conduction blocks along cardiac tissue walls. Another example of an energy delivery method is irreversible electroporation (IRE), which applies a high electric field that damages cell membranes. In a two-stage procedure (mapping followed by ablation), electrical activity at points within the heart 120 is typically sensed and measured by advancing a catheter 110 containing one or more electrical sensors (e.g., electrodes 111) into the heart 120 and acquiring / collecting data (e.g., biometric data generally or ECG data specifically) at multiple points. This data is then used to select a target region of the endocardium where ablation will be performed.

[0067] Cardiac ablations and other cardiac electrophysiology procedures are becoming increasingly complex as physicians treat challenging conditions such as atrial fibrillation and ventricular tachycardia. Treatment of refractory arrhythmias may currently rely on the use of three-dimensional (3D) mapping systems to reconstruct the anatomical structure of the targeted heart chamber. In this regard, the detection engine 101 employed by the system 100 herein manipulates and evaluates biometric data generally, and ECG data specifically, to generate improved tissue data that enables more accurate diagnoses, images, scans, and / or maps for treating abnormal heart rhythms or arrhythmias. For example, cardiologists rely on software such as the Complex Fractionated Atrial Electrograms (CFAE) module of the CARTO® 3 3D mapping system, manufactured by Biosense Webster, Inc. (Diamond Bar, Calif.), to generate and analyze ECG data. The detection engine 101 of the system 100 enhances this software to generate and analyze improved biometric data, which further provides multiple pieces of information regarding the electrophysiological properties of the heart 120 (including scar tissue) that are representative of the cardiac substrate (anatomical and functional) of the aFib.

[0068] Therefore, the system 100 can implement a 3D mapping system, such as the CARTO® 3 3D mapping system, to identify potential arrhythmogenic substrates of cardiomyopathies in terms of detecting abnormal ECGs. These cardiac disease-related substrates have been associated with the presence of split and prolonged ECGs in the endocardial and / or epicardial layers of the ventricular chambers (right and left). For example, low- or medium-voltage regions may indicate split and prolonged ECG activity. Furthermore, low- or medium-voltage regions during sinus rhythm may correspond to critical isthmuses identified during sustained, coherent ventricular arrhythmias (e.g., as applies to non-permissive ventricular tachycardia, as well as within the atria). Generally, abnormal tissue is characterized by low-voltage ECGs. However, early clinical experience with endocardial-epicardial mapping has shown that low-voltage regions are not always present as the sole arrhythmogenic mechanism in these patients. Indeed, areas of low or medium voltage may show ECG fragmentation and prolonged activity during sinus rhythm, which corresponds to the critical isthmus identified during sustained, coherent ventricular arrhythmias (e.g., only applicable to nonpermissive ventricular tachycardia). Furthermore, ECG fragmentation and prolonged activity are often observed in areas showing normal or near-normal voltage amplitudes (>1-1.5 mV). These latter areas can be evaluated according to voltage amplitude but cannot be considered normal according to the intracardiac signal and therefore represent true arrhythmogenic substrates. 3D mapping can identify the location of arrhythmogenic substrates on the endocardial and / or epicardial layers of the right and / or left ventricles, whose distribution may vary depending on the primary disease progression.

[0069] As another exemplary operation, cardiac mapping may be performed by the system 100 using one or more multi-electrode catheters (e.g., catheter 110). The multi-electrode catheter is used to stimulate and map electrical activity within the heart 120 and to ablate sites of abnormal electrical activity. In use, the multi-electrode catheter is inserted into a major vein or artery, such as the femoral vein, and then guided into a target chamber of the heart 120. A typical ablation procedure involves inserting the catheter 110, having at least one electrode 111 at its distal end, into the heart chamber. A reference electrode is provided by taping to the patient's skin, by a second catheter placed in or near the heart, or by selecting one or other of the electrodes 111 on the catheter 110. Radio frequency (RF) current is applied to the tip electrode 111 of the ablation catheter 110, causing current to flow in the medium surrounding the tip electrode (i.e., blood and tissue) toward the reference electrode. The distribution of the current depends on the amount of electrode surface in contact with tissue compared to blood, which has a higher electrical conductivity than tissue. Tissue heating occurs due to the electrical resistance of the tissue. Sufficient tissue heating induces cell destruction in the cardiac tissue, resulting in lesions within the non-conductive cardiac tissue. This process also heats the tip electrode 111 through conduction from the heated tissue to the electrode itself. If the electrode temperature becomes high enough, potentially exceeding 60°C, a thin, transparent film of dehydrated blood proteins can form on the surface of the electrode 111. As the temperature continues to rise, this dehydrated layer gradually thickens, causing blood to coagulate on the electrode surface. Because dehydrated biological material has a higher electrical resistance than endocardial tissue, the impedance to the flow of electrical energy into the tissue also increases. If the impedance becomes high enough, an impedance rise occurs, requiring the catheter 110 to be removed from the body and the tip electrode 111 to be cleaned.

[0070] Referring now to FIG. 2, a diagram of a system 200 capable of implementing one or more features of the subject matter of this disclosure is shown, according to one or more exemplary embodiments. System 200 includes, for a patient 202 (e.g., an example of patient 125 in FIG. 1 ), an apparatus 204, a local computing device 206, a remote computing system 208, a first network 210, and a second network 211. Additionally, apparatus 204 may include a biometric sensor 221 (e.g., an example of catheter 110 in FIG. 1 ), a processor 222, a user input (UI) sensor 223, a memory 224, and a transceiver 225. Note that for ease of explanation and brevity, detection engine 101 from FIG. 1 is reused in FIG. 2 .

[0071] According to one embodiment, device 204 may be an example of system 100 of FIG. 1 , where device 204 may include both patient-internal and patient-external components. According to another embodiment, device 204 may be a patient 202-external device including an attachable patch (e.g., attached to the patient's skin). According to another embodiment, device 204 may be internal to the body of patient 202 (e.g., subcutaneously implantable), where device 204 may be inserted into patient 202 by any applicable method, including oral infusion, surgical insertion via a vein or artery, endoscopic surgery, or laparoscopic surgery. According to one embodiment, although a single device 204 is shown in FIG. 2 , an exemplary system may include multiple devices.

[0072] Thus, the device 204, the local computing device 206, and / or the remote computing system 208 can be programmed to execute computer instructions related to the detection engine 101. As an example, the memory 223 stores these instructions for execution by the processor 222 such that the device 204 can receive and process biometric data via the biometric sensor 201. In this manner, the processor 222 and the memory 223 are representative of the processor and memory of the local computing device 206 and / or the remote computing system 208.

[0073] The apparatus 204, the local computing device 206, and / or the remote computing system 208 can be any combination of software and / or hardware that individually or collectively stores, executes, and implements the detection engine 101 and its functions. Furthermore, the apparatus 204, the local computing device 206, and / or the remote computing system 208 can be an electronic computer framework that includes and / or uses any number and combination of computing devices and networks utilizing various communication technologies, as described herein. The apparatus 204, the local computing device 206, and / or the remote computing system 208 can be easily scalable, extensible, and modular, capable of being tailored for different services or reconfigured with some functions independent of others.

[0074] Networks 210 and 211 may be wired networks, wireless networks, or may include one or more wired and wireless networks. According to one embodiment, network 210 is an example of a short-range network (e.g., a local area network (LAN) or a personal area network (PAN)). Information can be transmitted over short-range network 210 between apparatus 204 and local computing device 206 using any one of a variety of short-range wireless communication protocols, such as Bluetooth, Wi-Fi, Zigbee, Z-Wave, near field communications (NFC), Ultraband, Zigbee, or infrared (IR). Furthermore, network 211 is an example of one or more of an intranet, a local area network (LAN), a wide area network (WAN), a metropolitan area network (MAN), a direct connection or series of connections, a cellular telephone network, or any other network or medium capable of facilitating communication between local computing device 206 and remote computing system 208. Information can be transmitted over network 211 using any one of a variety of long-range wireless communication protocols (e.g., TCP / IP, HTTP, 3G, 4G / LTE, or 5G / New Radio). It should be noted that the wired connections of networks 210 and 211 can be implemented using Ethernet, Universal Serial Bus (USB), RJ-11, or any other wired connection, and the wireless connections can be implemented using Wi-Fi, WiMAX, Bluetooth, infrared, cellular networks, satellite communications, or any other wireless connection method.

[0075] In operation, device 204 may continuously or periodically acquire, monitor, store, process, and communicate biometric data related to patient 202 via network 210. Additionally, device 204, local computing device 206, and / or remote computing system 208 communicate via networks 210 and 211 (e.g., local computing device 206 may be configured as a gateway between device 204 and remote computing system 208). For example, device 204 may be an example of system 100 of FIG. 1 configured to communicate with local computing device 206 via network 210. Local computing device 206 may be, for example, a fixed / standalone device, a base station, a desktop / laptop computer, a smartphone, a smartwatch, a tablet, or any other device configured to communicate with other devices via networks 211 and 210. A remote computing system 208, implemented as a physical server on or connected to the network 211 or as a virtual server within a public cloud computing provider of the network 211 (e.g., Amazon Web Services (AWS)®), can be configured to communicate with the local computing device 206 over the network 211, thereby enabling biometric data related to the patient 202 to be communicated throughout the system 200.

[0076] The elements of device 204 are described below. Biometric sensor 221 may include, for example, one or more transducers configured to convert one or more environmental conditions into electrical signals so that different types of biometric data may be observed / obtained / acquired. For example, biometric sensor 221 may include one or more of an electrode (e.g., electrode 111 of FIG. 1 ), a temperature sensor (e.g., a thermocouple), a blood pressure sensor, a blood glucose sensor, a blood oxygen sensor, a pH sensor, an accelerometer, and a microphone.

[0077] In executing the detection engine 101, the processor 222 can be configured to receive, process, and manage biometric data acquired by the biometric sensor 221 and communicate the biometric data to the memory 224 for storage and / or across the network 210 via the transceiver 225. Biometric data from one or more other devices 204 may also be received by the processor 222 via the transceiver 225. Additionally, as described in more detail herein, the processor 222 can be configured to selectively respond to different tapping patterns (e.g., single tap or double tap) received from the UI sensor 223 such that different tasks of the patch (e.g., acquiring, storing, or transmitting data) are initiated based on the detected pattern. In some embodiments, the processor 222 can generate audible feedback regarding detecting the gesture.

[0078] The UI sensor 223 may be, for example, a piezoelectric or capacitive sensor configured to receive user input, such as a tap or touch. For example, the UI sensor 223 may be controlled to perform capacitive coupling in response to the patient 202 tapping or touching the surface of the device 204. Gesture recognition may be implemented via any one of a variety of capacitive types, such as resistive capacitive, surface capacitive, projected capacitive, surface acoustic wave, piezoelectric, and infrared touch. The capacitive sensor may be disposed over a small area or length of the surface, such that a tap or touch on the surface activates the monitoring device.

[0079] The memory 224 may be any non-transitory, tangible medium, such as magnetic, optical, or electronic memory (e.g., any suitable volatile and / or non-volatile memory, such as a random access memory or a hard disk drive). The memory 224 stores computer instructions that are executed by the processor 222.

[0080] The transceiver 225 may include a separate transmitter and a separate receiver, or the transceiver 225 may include a transmitter and receiver integrated into a single device.

[0081] During operation, the device 204 utilizing the detection engine 101 observes / acquires biometric data of the patient 202 via the biometric sensor 221, stores the biometric data in memory, and shares this biometric data throughout the system 200 via the transceiver 225. The detection engine 101 can then utilize models, algorithms (e.g., deep learning optimization), neural networks, machine learning, and / or artificial intelligence to generate and provide a mapping to the physician, reducing the processing load on the system 100 and transforming the operating system 100 into a more accurate mapping machine.

[0082] 3, a method 300 is illustrated in accordance with one or more embodiments. Method 300 generally illustrates one or more operations of a detection engine 101 that performs deep learning optimization to detect persistence factors of aFibs to be ablated to treat persistent aFib patients and classify velocity vector field images and raw data into persistence factors.

[0083] The method 300 begins at block 307, where the detection engine 101 detects one or more segments of the LAT (e.g., for a first excitation time). In this regard, input to the detection engine 101 includes biometric data in the form of an ECG data signal. From the ECG data signal, one or more of the LATs are determined over at least one time interval (e.g., a segment). Specifically, the detection engine 101 determines a time point of threshold activity corresponding to local excitation, calculated based on a normalized initial starting point (e.g., for a first excitation time).

[0084] In block 315, the detection engine 101 calculates the field, such as by modeling a vector velocity field. For example, a vector velocity field that measures and quantifies the instantaneous velocity of the ECG data signal as it passes through the LAT (e.g., by calculating the direction of the wave at each x, y point and using the derivative of the polynomial surface to provide the velocity vector field). According to one or more embodiments, the catheter 110 can be positioned on the surface of the atrial (x, y) plane, and the detection engine 101 can describe the LAT using a scatter plot (e.g., with respect to the first operating time). The detection engine 101 can estimate the coefficients of the scatter plot and determine their derivatives to provide the velocity vector field.

[0085] In block 321, the detection engine 101 detects and classifies persistence factors. In this regard, the detection engine 101 uses the velocity vector field of block 315 as input to a machine learning and / or artificial intelligence algorithm (e.g., a deep convolutional neural network) to detect the location of a "gold standard" persistence factor. One or more advantages, technical effects, and / or benefits of block 321 include the exploitation of big data to provide an understanding of the outcome of a particular case, including whether the ablation outcome was successful (e.g., whether and to what extent the ablation procedure had one or more positive and negative outcomes). Thus, the detection engine 101 provides automatic understanding from a set of ECG signals that can be displayed via a GUI. The detection engine 101 performing method 300 can provide retrospective analysis by reviewing past cases to determine where ablation occurred, and can also provide predictive analysis to determine where to perform ablation in future cases.

[0086] According to one or more embodiments, the detection engine 101 uses a machine learning algorithm, such as a neural network, described herein, to determine cues in the ECG data signal regarding outcome (e.g., which portions of the data indicate when and whether the procedure had a positive outcome). Cues may also include using inputs such as, but not limited to, system state, ablation parameters, ablation location, ablation duration, applied force, power, and temperature. For example, after an ablation results in a change and maintenance of tachycardia, including prolongation and subsequent breakdown of the tachycardia, the machine learning algorithm of the detection engine 101 can automatically tag the event and / or location that led to termination. The machine learning algorithm of the detection engine 101 can further tag locations and events provided by the physician that indicate the ablation that led to termination, as predicted / expected by the machine learning algorithm. Thus, the detection engine 101 provides the user with an understanding of the clinical outcome of the electrophysiology procedure. Additionally, the operation of the detection engine 101 can be applied to other features of interest, such as HIS bundle (e.g., including special localization signals) and diaphragm capture (e.g., including data flag stimulation during AI sessions), and understanding the appearance of successful termination.

[0087] In connection with block 321 of method 300, FIG. 4 shows a diagrammatic depiction of an artificial intelligence system 400 according to one or more embodiments. The artificial intelligence system 400 includes data 410 (e.g., biometric data), a machine 420, a model 430, an outcome 440, and (underlying) hardware 450. Where appropriate, for ease of understanding, the description of FIGS. 4-5 will be made with reference to FIGS. 1-3. For example, the machine 410, the model 430, and the hardware 450 may represent aspects of the detection engine 101 of FIGS. 1-2 (e.g., the machine learning and / or artificial intelligence algorithms therein), while the hardware 450 may also represent the catheter 110 of FIG. 1, the console 160 of FIG. 1, and / or the device 204 of FIG. 2. Generally, the machine learning and / or artificial intelligence algorithms of the artificial intelligence system 400 (e.g., implemented by the detection engine 101 of FIGS. 1-2 ) operate on the hardware 450 using the data 410 to train a machine 420 to build a model 430 and predict an outcome 440.

[0088] For example, machine 420 may act as a controller or data collection associated with hardware 450 and / or may be associated with hardware 450. Data 410 (e.g., biometric data as described herein) may be ongoing data or output data associated with hardware 450. Data 410 may also include currently collected data, historical data, or other data from hardware 450, may include measurements during a surgical procedure and may be associated with the outcome of the surgical procedure, may include temperatures of heart 140 of FIG. 1 that were collected and correlated with the outcome of the cardiac procedure, and may be associated with hardware 450. Data 410 may be divided into one or more subsets by machine 420.

[0089] Additionally, the machine 420 is trained, such as against the hardware 450. This training may also include analysis and correlation of the collected data 410. According to one or more embodiments, the detection engine 101 may train machine learning algorithms to determine the termination of an acute arrhythmia, the termination of an aFib, or the termination of any tachycardia, and / or to identify outcomes after a blanking period of a certain number of days and inform the use of long-term follow-up that may be used.

[0090] For example, in the cardiac case, temperature and outcome data 410 may be trained to determine whether a correlation or association exists between the temperature of the heart 140 of Figure 1 and the outcome during a cardiac procedure. According to another embodiment, training the machine 420 may include self-training by the detection engine 101 of Figure 1 with one or more subsets. In this regard, the detection engine 101 of Figure 1 learns to detect point-by-point case classifications.

[0091] Additionally, model 430 is constructed on data 410 associated with hardware 450. Constructing model 430 can include physical hardware or software modeling, algorithmic modeling, and / or similar modeling intended to represent collected and trained data 410 (or a subset thereof). In some aspects, constructing model 430 is part of a self-training operation by machine 420. Model 430 can be configured to model the operation of hardware 450 and model data 410 collected from hardware 450 to predict outcome 440 obtained by hardware 450. Predicting outcome 440 (of model 430 associated with hardware 450) can use trained model 430. By way of example and to further understand the present disclosure, for the heart, if temperatures during treatment between 36.5°C and 37.89°C (i.e., 97.7°F and 100.2°F) result in positive outcomes from cardiac treatments, outcome 440 can be predicted for a particular treatment using these temperatures. Thus, the predicted outcome 440 can be used to configure the machine 420, model 430, and hardware 450 accordingly.

[0092] Thus, the machine learning and / or artificial intelligence algorithms therein may include neural networks, in order for the artificial intelligence system 400 to use data 410 to operate against hardware 450 to train machines 420, build models 430, and predict outcomes 440. Generally, a neural network is a network or circuit of neurons, or in the modern sense, an artificial neural network (ANN) made up of artificial neurons or nodes or cells.

[0093] For example, an ANN contains a network of processing elements (artificial neurons) that can exhibit complex global behavior determined by the connections between the processing elements and element parameters. These connections in a neuronal network or circuit are modeled as weights. Positive weights reflect excitatory connections, while negative values ​​represent inhibitory connections. Inputs are modified by the weights and summed using linear combinations. An activation function can control the amplitude of the output. For example, the acceptable range of the output is typically 0 to 1, or sometimes -1 to 1. ANNs are often adaptive systems that change their structure based on external or internal information flowing through the network.

[0094] In more practical terms, neural networks are nonlinear statistical data modeling or decision-making tools that can be used to model complex relationships between inputs and outputs or to find patterns in data. Therefore, ANNs can be used for predictive modeling and adaptive control applications while being trained through a data set. Note that self-learning arising from experience can occur within ANNs, allowing them to draw conclusions from complex and seemingly unrelated sets of information. The usefulness of artificial neural network models lies in the fact that they can be used to estimate and use functions from observations. Unsupervised neural networks can also be used to learn representations of inputs that capture salient features of the input distribution, and more recently, deep learning algorithms that can implicitly learn distribution functions for observed data. Training with neural networks is particularly useful in applications where the complexity of the data (e.g., biometric data) or the task (e.g., monitoring, diagnosing, and treating any number of different diseases) makes the design of such functions impractical.

[0095] Neural networks can be used in a variety of fields. Thus, in the artificial intelligence system 400, the machine learning and / or artificial intelligence algorithms therein can include neural networks broadly divided according to the tasks to which they are applied. These divisions tend to fall into the following categories: regression analysis (e.g., function approximation), including time series prediction and modeling; classification, including pattern and sequence recognition; novelty detection and continuous decision-making; data processing, including filtering; clustering; blind source separation; and compression. For example, application areas of ANNs include nonlinear system identification and control (vehicle control, process control), game playing and decision-making (backgammon, chess, racing), pattern recognition (radar systems, face identification, object recognition), sequence recognition (gesture, speech, handwritten text recognition), medical diagnosis and treatment, financial applications, data mining (or knowledge discovery in databases, or "KDD"), visualization, and email spam filtering. For example, it is possible to create semantic profiles of patient biometric data obtained from medical procedures.

[0096] According to one or more embodiments, the neural network may implement a long-short-term memory neural network architecture, a convolutional neural network (CNN) architecture, a recurrent neural network (RNN) architecture, etc. The neural network may include multiple layers, multiple connections (e.g., encoder-decoder connections), regularization methods (e.g., dropout), and optimization properties.

[0097] Long-short-term memory neural network architectures contain feedback connections and can process single data points (e.g., images) along with entire sequences of data (e.g., speech or video). The units of a long-short-term memory neural network architecture can consist of cells, input gates, output gates, and forget gates, where cells store values ​​over any time interval and gates regulate the flow of information to and from cells.

[0098] The CNN architecture is a shared weight architecture with translational invariance properties, where each neuron in one layer is connected to every neuron in the next layer. The regularization technique of the CNN architecture can exploit hierarchical patterns in the data and organize more complex patterns using smaller and simpler patterns. When a neural network implements a CNN architecture, other configurable aspects of the architecture include the number of filters at each stage, the kernel size, and the number of kernels per layer.

[0099] Referring now to FIG. 5, a block diagram of an example neural network 500 and a method 501 performed within the neural network 500 is shown in accordance with one or more embodiments. The neural network 500 operates to assist in the implementation of the machine learning and / or artificial intelligence algorithms described herein (e.g., as implemented by the detection engine 101 of FIGS. 1-2). The neural network 500 may be implemented in hardware, such as the machine 420 and / or hardware 450 of FIG. 4. As provided herein, the description of FIGS. 4-5 will be made with reference to FIGS. 1-3, where appropriate, for ease of understanding.

[0100] In an example operation, the detection engine 101 of FIG. 1 includes collecting data 410 from hardware 450. In the neural network 500, the input layer 510 is represented by multiple inputs (e.g., inputs 512 and 514 of FIG. 5 ). With reference to block 520 of the method 501, the input layer 510 receives the inputs 512 and 514. The inputs 512 and 514 may include biometric data. For example, collecting the data 410 may be aggregating biometric data (e.g., BS ECG data, IC ECG data, and ablation data, along with catheter electrode position data) from one or more treatment records of the hardware 450 into a dataset (represented by data 410).

[0101] At block 525 of method 501, neural network 500 encodes inputs 512 and 514 using any portion of data 410 (e.g., datasets and predictions generated by artificial intelligence system 400) to generate latent representations or data encodings. The latent representations include one or more intermediate data representations derived from multiple inputs. According to one or more embodiments, the latent representations are generated by element-by-element activation functions (e.g., sigmoid functions or rectified linear functions) of detection engine 101 of FIG. 1. As shown in FIG. 5, inputs 512 and 514 are provided to hidden layer 530, which is shown to include nodes 532, 534, 536, and 538. Neural network 500 executes processing through hidden layer 530 of nodes 532, 534, 536, and 538 to exhibit complex global behavior determined by the connections between processing elements and element parameters. Thus, the transition between layers 510 and 530 can be viewed as an encoder stage that takes inputs 512 and 514 and forwards them to the deep neural network (in layer 530) to learn some smaller representations (e.g., the resulting latent representations).

[0102] The deep neural network may be a CNN, a long-short-term memory neural network, a fully connected neural network, or a combination thereof. The inputs 512 and 514 may be intracardiac ECG, a surface ECG, or both intracardiac ECG and surface ECG. This encoding results in dimensionality reduction of the inputs 512 and 514. Dimensionality reduction is the process of reducing the number of random variables (in the inputs 512 and 514) being considered by obtaining a set of key variables. For example, dimensionality reduction may be feature extraction, which transforms data (e.g., inputs 512 and 514) from a high-dimensional space (e.g., greater than 10 dimensions) to a low-dimensional space (e.g., 2-3 dimensions). Technical effects and advantages of dimensionality reduction include reducing the time and storage space requirements of the data 410, improving visualization of the data 410, and improving parameter interpretability for machine learning. This data transformation may be linear or nonlinear. The receiving (block 520) and encoding (block 525) operations can be considered the data preparation portion of a multi-stage data manipulation by the detection engine 101.

[0103] At block 545 of method 510, neural network 500 decodes the latent representation. The decoding stage receives the encoder output (e.g., the resulting latent representation) and attempts to reconstruct some form of inputs 512 and 514 using another deep neural network. In this regard, nodes 532, 534, 536, and 538 generate output 552 in output layer 550, as shown at block 560 of method 510. That is, output layer 550 reconstructs inputs 512 and 514 with reduced dimensionality, but without signal interference, signal artifacts, and signal noise. An example of output 552 includes cleaned biometric data (e.g., a clean / denoised version of IC ECG data). Technical effects and benefits of cleaned biometric data include enabling more accurate monitoring, diagnosis, and treatment of any number of various disorders.

[0104] 3, method 300 continues at block 330, where detection engine 101 determines / calculates a code for each point in the (x,y) plane for one of four directions (e.g., left can equal red, right can equal green, up can equal blue, and down can equal yellow) to provide a color-coded vector field image, which can enable localization of focal excitation sources or sustained sites.

[0105] In block 345, the detection engine 101 detects focal activation / rotor indicators, for example, by scanning the color-coded vector field image of block 330 using a kernel (e.g., a circle, typically 1 mm in radius). Generally, the kernel can be a specified space within the map and can be of a desired size (e.g., a circle or square) and dimensions. In some cases, the kernel can be the address of each point within the map, a calculated section of the map, a calculated boundary of the map, etc. In cardiac tissue, swirling wave reentry occurs when a wavefront of electrical propagation encounters functionally non-excitable tissue and rotates around it in a spiral fashion. Similarly, a rotor indicator or rotor can be a center of rotation from which two-dimensional swirling waves of activation rotate outward. Furthermore, a focal activation indicator or focal activation can be an arrhythmia with an electrical impulse originating within and confined within the atrium. Focal activation / rotor indicators can be detected within the kernel and when all directions are sequential (e.g., clockwise or counterclockwise). Thus, the detection engine 101 distinguishes between focal excitation sources and rotors (eg, active and inactive) on the modeling vector velocity field.

[0106] At block 360, the detection engine 101 marks persistence factors. According to one or more embodiments, persistence factors are tissue triggers and / or initiating factors that ensure the persistence of aFib. The detection engine 101 can automatically identify or annotate aFib persistence factors based on vector velocity and ablation information. More specifically, the detection engine 101 can determine one or more gold standard annotations (e.g., most accurate annotations) of aFib persistence factors. Alternatively, or in combination with the above, a physician can mark suspected persistence factors that may be ablated in an ablation procedure.

[0107] In block 375, the detection engine 101 classifies the focal excitation / rotor indices into sustained factors. Classifying the focal excitation / rotor indices into sustained factors includes classifying active or inactive regions of interest (ROIs) based on the optical properties and the vector field map using machine learning and / or artificial intelligence (e.g., logistic regression classifiers, support vector machines, deep learning, etc.). For example, a vector velocity image with gold standard annotations of sustained factors can be used as input and / or target for a neural network (e.g., an RNN or CNN described herein) to predict whether a pixel in the vector velocity image is a sustained factor. Furthermore, classifying one or more focal excitation and / or rotor indices into sustained factors can use the velocity vector field as input to a neural network (e.g., an RNN or CNN described herein) used to detect the location of the gold standard sustained factor.

[0108] According to one or more embodiments, ROI annotations can be used by the detection engine 101 for active, inactive, and unknown categories. For example, an active ROI can indicate that the ROI was ablated, resulting in the termination of aFib, or that the ablation session increased cycle length (CL). An active ROI includes a persistence factor, which has "clinical value" including evidence that an ablation session near a focal activation / rotor resulted in an increase in cycle length or the termination of atrial fibrillation. An inactive ROI can indicate no visible change in aFib characteristics, including an ablation session near a focal activation source without a visible change in atrial fibrillation characteristics. An unknown ROI annotation can indicate no visible ablation session near a focal activation / rotor. In this manner, one or more advantages, technical effects, and / or benefits of the detection engine include addressing most points in the mapping that are not addressed by a physician because the active or inactive nature of the ROI is unknown to the physician. The detection engine 101 may further assign probabilities across active, inactive, and unknown categories (eg, a scale of 0 to 100 indicating the likelihood of a category).

[0109] With respect to the operations of blocks 321, 360, and 375, it should be noted that the detector engine 101 may use the results of any operation as feedback or input to another operation, as indicated by the double-headed arrow.

[0110] 6, a method 600 is illustrated in accordance with one or more embodiments. Method 600 generally illustrates one or more operations of detection engine 101 that implements optimization to detect perpetrators of atrial fibrillation that are to be ablated for treatment.

[0111] At block 605, the method 600 begins, where the detection engine 101 receives one or more inputs, such as an IC ECG data signal from the catheter 110. At block 610, the detection engine 101 detects a segment of the LAT. Figure 7 illustrates a graph 700 in accordance with one or more embodiments. As shown in Figure 7, a segment of the LAT is detected for an excitation time relative to a first excitation time (shown as a circle 710).

[0112] In block 625, the detection engine 101 determines / calculates / models a vector velocity field. In this regard, the detection engine 101 assumes that the catheter 110 is located on a surface in the atrial (x,y) plane and uses a scatter plot to describe the time of activation (e.g., for the first activation time depicted as a circle in FIG. 7). FIG. 8 shows a graph 800 of a surface (x,y) in accordance with one or more embodiments. The graph 800 provides a plot including the surface (x,y) and Z (milliseconds) of the activation time within a segment. Based on the graph 800, the detection engine 101 can estimate the coefficients of the surface (x,y) that best fit the dot. The polynomial surface T(x,y) is calculated by fitting N and a to the cost function (see Equation 1) using the steepest descent method. i,j can be estimated, and the position x s , y s The local excitation time L measured in (x s ,y s ) plane, the estimated surface T(x s ,y s ) is minimized by a regularization term (ρ typically equal to 0.1) that takes into account the number of parameters estimated in the model. More specifically, the polynomial surface can be defined by Equation 2:

[0113]

number

[0114] From T(x,y) to the vector field, the model can be used to calculate the direction of the electric wave at each point in (x,y) defined by Equation 3.

[0115]

number

[0116] T x is defined by Equation 4.

[0117]

number

[0118] Figure 9 shows a graph 900 of the velocity vector field of the surface (x,y) of Figure 8, in accordance with one or more embodiments. That is, the derivative of the polynomial surface of Figure 8 gives the velocity vector field shown in Figure 9.

[0119] In block 630, each point in the (x,y) plane is coded into one of four directions to give a color-coded vector field image, which allows for the localization of focal excitation sources or sustainers.

[0120] In block 645, foci / rotors can be detected by the detection engine 101. For example, the detection engine 101 can use a kernel (typically a circle with a radius of 1 mm) to scan the color-coded vector field image acquired in block 630. A rotor / foci point is detected if all four directions appear in sequence (clockwise or counterclockwise) within the kernel.

[0121] In block 660, the persistence factor may be marked by the detection engine 101. In block 675, the focal excitation / rotor indicator may be classified by the detection engine 101 as a persistence factor.

[0122] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of instructions, which includes one or more executable instructions for implementing the depicted logical function(s). In some alternative implementations, the functions shown in the blocks may occur out of the order shown in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. It should also be noted that each block of the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may operate or execute a combination of dedicated hardware and computer instructions.

[0123] While features and elements are described above in particular combinations, those skilled in the art will understand that each feature or element can be used alone or in combination with other features and elements. Additionally, the methods described herein may be implemented in a computer program, software, or firmware embodied in a computer-readable medium for execution on a computer or processor. As used herein, computer-readable medium should not be construed as being a transitory signal itself, such as an electric wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse passing through a fiber optic cable), or an electrical signal transmitted through a current line.

[0124] Examples of computer-readable media include electrical signals (transmitted via wired or wireless connections) and computer-readable storage media. Examples of computer-readable storage media include, but are not limited to, registers, cache memory, semiconductor memory devices, magnetic media such as internal hard disks and removable disks, magneto-optical media, optical media such as compact discs (CDs) and digital versatile discs (DVDs), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), and memory sticks. A processor together with software can implement a radio frequency transceiver for use in a terminal, a base station, or any host computer.

[0125] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly dictates otherwise. It should be understood that the terms "comprise" and / or "comprising," as used herein, indicate the presence of stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0126] The description of different embodiments herein is provided for illustrative purposes, but is not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terms used herein are selected to best explain the principles, practical applications, or technical improvements of the embodiments compared to technologies found on the market, or to enable those skilled in the art to understand the embodiments disclosed herein.

[0127] [Embodiment] (1) A method implemented by a detection engine embodied as processor-executable code stored on a memory and executed by at least one processor, comprising: modeling, by said detection engine, a vector velocity field that measures and quantifies the velocity of electrocardiogram data signals through local activation times; determining, by said detection engine, one or more codes for each point in the plane to provide a color-coded vector field image; detecting, by the detection engine, focal and rotor indications by scanning the color-coded vector field image using one or more kernels; and classifying, by the detection engine, the focal excitation and rotor indicators into perpetuators. (2) The method of embodiment 1, wherein the electrocardiogram data signal is detected by a catheter within an anatomical structure in communication with the detection engine. (3) The method of embodiment 1, wherein the detection engine detects one or more segments of the local excitation time relative to a first excitation time. (4) The method of embodiment 1, wherein the detection engine calculates the direction of the radio waves at each x, y point and models a vector velocity field by using derivatives of a polynomial surface to provide the velocity vector field. (5) The method of embodiment 1, wherein the detection engine classifies the persistence factor by using the velocity vector field as input to a machine learning or artificial intelligence algorithm.

[0128] (6) The method of embodiment 5, wherein the machine learning or artificial intelligence algorithm comprises a deep convolutional neural network or a recurrent neural network for detecting the location of a gold standard persistence factor among the persistence factors. (7) The method of embodiment 5, wherein the machine learning or artificial intelligence algorithm determines whether the ablation outcome for a particular case related to the electrocardiogram data signal was successful. (8) The method of embodiment 1, wherein the focal excitation and rotor indices are detected when all directions are in order within the one or more kernels. (9) The method of embodiment 1, wherein the detection engine automatically identifies and annotates atrial fibrillation persistence factors based on vector velocity and ablation information when marking the persistence factors. (10) The method of embodiment 1, wherein annotations of the region of interest are used by the detection engine with respect to active, inactive, and unknown categories to indicate at least whether the region of interest has been ablated or whether atrial fibrillation has terminated.

[0129] (11) A system comprising: a memory storing processor-executable code for a detection engine; at least one processor executing said processor executable code to provide said system with: modeling, by said detection engine, a vector velocity field that measures and quantifies the velocity of electrocardiogram data signals through local activation times; determining, by said detection engine, one or more codes for each point in the plane to provide a color-coded vector field image; detecting, by the detection engine, focal excitation and rotor indices by scanning the color-coded vector field image using one or more kernels; and at least one processor that causes the detection engine to classify the focal excitation and rotor indicators into sustained factors. (12) The system of embodiment 11, wherein the electrocardiogram data signal is detected by a catheter within the anatomical structure in communication with the detection engine. (13) The system of embodiment 11, wherein the detection engine detects one or more segments of the local excitation time relative to a first excitation time. (14) The system of embodiment 11, wherein the detection engine calculates the direction of the radio waves at each x, y point and models a vector velocity field by using derivatives of a polynomial surface to provide the velocity vector field. (15) The system of embodiment 11, wherein the detection engine classifies the persistence factor by using the velocity vector field as input to a machine learning or artificial intelligence algorithm.

[0130] (16) The system of claim 15, wherein the machine learning or artificial intelligence algorithm comprises a deep convolutional neural network or a recurrent neural network for detecting the location of a gold standard persistence factor among the persistence factors. (17) The system of embodiment 15, wherein the machine learning or artificial intelligence algorithm determines whether the ablation outcome for a particular case related to the electrocardiogram data signal was successful. (18) The method of embodiment 11, wherein the focal excitation and rotor indices are detected when all directions are in order within the one or more kernels. (19) The system of embodiment 11, wherein the detection engine automatically identifies and annotates atrial fibrillation persistence factors based on vector velocity and ablation information when marking the persistence factors. (20) The system of embodiment 11, wherein annotations of the region of interest are used by the detection engine with respect to active, inactive, and unknown categories to indicate at least whether the region of interest has been ablated or whether atrial fibrillation has terminated.

Claims

1. A method of operating a system having a memory storing processor-executable code for a detection engine and a processor that executes the processor-executable code, comprising: modeling, by the processor, a vector velocity field that measures and quantifies the velocity of the electrocardiogram data signal through a local activation time; determining, by the detection engine, one or more codes for each point in the plane to provide a color-coded vector field image; detecting, by the processor, focal excitation and rotor indices by scanning the color-coded vector field image using one or more kernels; and classifying, by the processor, the focal excitation and rotor indices into sustained factors.

2. The method of claim 1 , wherein the electrocardiogram data signals are detected by a catheter within the anatomy in communication with the detection engine.

3. An operating method as described in claim 1, wherein the processor detects, by the detection engine, one or more segments of the local excitation time relative to a first excitation time.

4. The method of operation described in claim 1, wherein the processor calculates the direction of the radio waves at each x, y point using the detection engine and models a vector velocity field by using derivatives of a polynomial surface to provide the vector velocity field.

5. The operating method of claim 1, wherein the processor classifies the persistence factors by using the detection engine to use the vector velocity field as input to a machine learning or artificial intelligence algorithm.

6. 6. The method of claim 5, wherein the machine learning or artificial intelligence algorithm comprises a deep convolutional neural network or a recurrent neural network for detecting the location of gold standard persistence factors among the persistence factors.

7. The method of claim 5 , wherein the machine learning or artificial intelligence algorithm determines whether the ablation outcome for a particular case related to the electrocardiogram data signal was successful.

8. The method of claim 1 , wherein the focal excitation and rotor indices are detected when all directions are in order within the one or more kernels.

9. The operating method described in claim 1, wherein when the processor marks the persistence factors using the detection engine, the processor automatically identifies and annotates atrial fibrillation persistence factors based on vector velocity and ablation information.

10. 2. The method of claim 1, wherein annotations of a region of interest are used by the detection engine with respect to active, inactive, and unknown categories to indicate at least whether the region of interest has been ablated or whether atrial fibrillation has been terminated.

11. 1. A system comprising: a memory storing processor-executable code for a detection engine; at least one processor executing the processor executable code to provide the system with: modeling, by said detection engine, a vector velocity field that measures and quantifies the velocity of electrocardiogram data signals through local activation times; determining, by the detection engine, one or more codes for each point in the plane to provide a color-coded vector field image; detecting, by the detection engine, focal excitation and rotor indices by scanning the color-coded vector field image using one or more kernels; and at least one processor that causes the detection engine to classify the focal excitation and rotor indices into sustained factors.

12. The system of claim 11 , wherein the electrocardiogram data signals are detected by a catheter within the anatomy in communication with the detection engine.

13. The system of claim 11 , wherein the detection engine detects one or more segments of the local excitation time relative to a first excitation time.

14. 12. The system of claim 11, wherein the detection engine calculates the direction of the radio wave at each x,y point and models a vector velocity field by using derivatives of a polynomial surface to provide the vector velocity field.

15. The system of claim 11 , wherein the detection engine classifies the persistence factors by using the vector velocity field as input to a machine learning or artificial intelligence algorithm.

16. 16. The system of claim 15, wherein the machine learning or artificial intelligence algorithm comprises a deep convolutional neural network or a recurrent neural network for detecting the location of gold standard persistence factors among the persistence factors.

17. 16. The system of claim 15, wherein the machine learning or artificial intelligence algorithm determines whether ablation outcome for a particular case related to the electrocardiogram data signal was successful.

18. The method of claim 11 , wherein the focal excitation and rotor indices are detected when all directions are in order within the one or more kernels.

19. The system of claim 11 , wherein the detection engine automatically identifies and annotates atrial fibrillation persistence factors based on vector velocity and ablation information when marking the persistence factors.

20. 12. The system of claim 11, wherein annotations of a region of interest are used by the detection engine with respect to active, inactive, and unknown categories to indicate at least whether the region of interest has been ablated or whether atrial fibrillation has been terminated.

Images