Systems and methods for improving cardiac ablation procedures

Abstract

Systems and methods for improving cardiac ablation procedures are disclosed. The systems and methods include a cloud server including a database containing information about previously performed cardiac ablations, a local server communicably coupled to the cloud server via a first network, and a surgical system communicably coupled to the local server via a second network. The cloud server is configured to receive electrical data and anatomical data of a patient's heart from the surgical system via the local server, perform a comparison of the electrical data and the anatomical data of the heart to the database information, generate a surgical treatment plan for the heart based on the comparison, and transmit the surgical treatment plan to the surgical system via the local server, wherein the surgical treatment plan includes a mesh of the heart.

Description

TECHNICAL FIELD

[0001] The present application provides systems, devices, and methods for improving cardiac ablation procedures. BACKGROUND

[0002] Cardiac ablation is a surgical procedure to treat abnormal heart rhythms in a patient by burning or destroying tissue in the patient's heart. Cardiac ablation is often used to treat arrhythmias, such as atrial fibrillation, atrial flutter, supraventricular tachycardia, and Wolff-Parkinson-White syndrome. Cardiac ablation can prevent the abnormal heart rhythm from moving through the heart. In a cardiac ablation procedure, electrical measurements of the heart are made using a catheter. Based on the measurements, the surgeon uses heat (radiofrequency), extreme cold (cryoablation), or lasers to destroy the areas of the heart where the electrical abnormalities occur.

[0003] Traditional cardiac ablation procedures rely only on the subjective skills of the surgeon to determine where and how to perform the ablation. This traditional approach results in high variability in clinical outcomes, including high variability between patients, between surgeons, and between hospitals. BRIEF DESCRIPTION OF DRAWINGS

[0004] A more detailed understanding can be had from the following description, given by way of example in conjunction with the accompanying drawings wherein:

[0005] Figure 1 is an illustration of an example system in which one or more features of the disclosure can be implemented;

[0006] Figure 2A a process of performing analysis of cardiac ablation data is shown;

[0007] Figure 2B a process for generating a cardiac ablation treatment plan according to certain embodiments is shown;

[0008] Figure 3 a process for performing a cardiac ablation treatment according to certain embodiments is shown;

[0009] Figure 4 a graphical depiction of an artificial intelligence system is shown;

[0010] Figure 5 a method performed in an artificial intelligence system is shown; Figure 4

[0011] Figure 6 an example of a probability of a Naive Bayes calculation is shown;

[0012] Figure 7 an example decision tree is shown;

[0013] Figure 8 ​An example random forest classifier is shown;

[0014] Figure 9 An example logistic regression is shown;

[0015] Figure 10 An example support vector machine is shown;

[0016] Figure 11 An example linear regression model is shown;

[0017] Figure 12 An example K-means clustering is shown;

[0018] Figure 13 An example ensemble learning algorithm is shown;

[0019] Figure 14 An example neural network is shown;

[0020] Figure 15 A hardware-based neural network is shown;

[0021] Figure 16 is an illustration of an example system that can implement one or more features of the disclosed subject matter;

[0022] Figure 17 A process for scoring a cardiac ablation therapy is shown, according to certain embodiments;

[0023] Figure 18 A process for training a surgeon to perform cardiac ablation using an AR / VR system is shown, according to certain embodiments;

[0024] Figure 19A An example of a map that can be generated by a surgical system is shown;

[0025] Figure 19B An example of a map that can be generated by a surgical system is shown;

[0026] Figure 19C Another example of a map that can be generated by a surgical system is shown;

[0027] Figure 20A A display of RF index Shure points is shown;

[0028] Figure 20B A display of time-based identification of ablation points is shown;

[0029] Figure 20C A display is shown that shows a stability view of an RF index detail representation.

[0030] Figure 20D and Figure 20EAn additional display showing the display of additional parameters to the using physician is shown.

[0031] Figure 21A is a graphical representation of a first modified cardiac ablation treatment plan;

[0032] Figure 21B is a graphical representation of a location of a first ablation performed;

[0033] Figure 21C is a graphical representation of a first modified cardiac ablation treatment plan;

[0034] Figure 21D is a graphical representation of a location of a second ablation performed;

[0035] Figure 21E is a graphical representation of a second modified cardiac ablation treatment plan;

[0036] Figure 21F is a graphical representation of a location of a third ablation performed;

[0037] Figure 21G is a graphical representation of a third modified cardiac ablation treatment plan; and

[0038] Figure 21H is a graphical representation of a location of a fourth ablation performed. DETAILED DESCRIPTION

[0039] The present invention utilizes advances in machine learning and big data analysis to improve the clinical technique of cardiac ablation procedures. Embodiments of the present invention can use data collected from multiple patients, multiple surgeons, and multiple hospitals to generate a cardiac ablation treatment plan. Embodiments of the present invention further improve clinical outcomes by providing systems and methods that dynamically adjust while performing a cardiac ablation procedure. Embodiments of the present invention are able to achieve improved clinical outcomes while also remaining in compliance with the Health Insurance Portability and Accountability Act of 1996 (HIPAA) and the General Data Protection Regulation (GDPR) of the European Union.

[0040] The present invention includes a system and method for improving a cardiac ablation procedure. The system and method include a cloud server including a database containing information about previously performed cardiac ablations, a local server communicably coupled to the cloud server via a first network, and a surgical system communicably coupled to the local server via a second network, wherein the cloud server is configured to receive electrical data and anatomical data of a patient’s heart from the surgical system via the local server, perform a comparison of the electrical data and anatomical data of the heart to the database information, generate a surgical treatment plan for the heart based on the comparison (wherein the surgical treatment plan includes a mesh of the heart), and transmit the surgical treatment plan to the surgical system via the local server. The system and method can include changing one or more parameters of the surgical system in accordance with the treatment plan. The cloud server can receive electrical data and anatomical data of a patient’s heart from the surgical system in performing a cardiac ablation procedure. The cloud server can also be configured to receive additional electrical data and anatomical data of the heart after performing the ablation procedure and generate an ablation score based on the additional electrical data and anatomical data. The ablation procedure can include a plurality of ablations, wherein the cloud server receives electrical data and anatomical data of a patient’s heart from the surgical system after each ablation.

[0041] The present invention includes a system and method for training a surgeon to perform a cardiac ablation. The system and method include a cloud server containing database information about previously performed cardiac ablations, and an AR / VR system communicably coupled to the cloud server via a first network, wherein the AR / VR system is configured to receive input defining parameters of a training ablation, receive electrical data and anatomical data for a virtual patient from the cloud server (wherein the electrical data and anatomical data for the virtual patient are determined by comparing the input to the database information), create a virtual simulation of a patient’s heart based on the electrical data and anatomical data for the virtual patient, measure a trainee’s performance in performing a cardiac ablation procedure on the virtual simulation, and score the trainee’s performance.

[0042] Figure 1 is an illustration of an example system 100 in which one or more features of the present disclosure can be implemented. In the system 100, a plurality of discrete networks (denoted as discrete surgical networks 101A-101N) are connected to a cloud-based platform 160 over a common network 150. In some cases, the cloud-based platform 160 is implemented by a public cloud computing platform (such as, for example, Amazon Web Services or Microsoft Azure), a hybrid cloud computing platform (such as, for example, HP Enterprise OneSphere), or a private cloud computing platform.

[0043] Discrete networks 101A-101N can be located within a single physical location, within a single entity network, on a physical or entity boundary, such as, for example, in a single hospital or a single healthcare provider network.

[0044] In one embodiment, each of the discrete networks 101 includes one or more surgical systems 110 connected to a local server 120. The one or more surgical systems 110 are capable of obtaining anatomical and electrical measurements of a patient’s heart and performing a cardiac ablation procedure. An example of a surgical system that can be used in system 100 is the Carto(R) 3 system sold by Biosense Webster(R). In some cases, the surgical systems 110 can also associate measurements with a unique patient identity (ID) or other information that can be used to uniquely identify a patient.

[0045] The surgical systems 110 can also, and optionally, use ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), or other medical imaging techniques known in the art to obtain anatomical measurements of a patient’s heart or other measurements related to the patient. The surgical systems 110 can use catheters, electrocardiograms (EKGs), or other sensors that measure electrical properties of the heart to obtain electrical measurements. The anatomical measurements and electrical measurements can then be stored in a local memory of the surgical systems 110 and transmitted to the local server 120 using the private network 105. In some cases, the electrical measurements and anatomical measurements are transmitted to the local server 120 immediately upon acquisition.

[0046] The surgical systems 110 then generate a map of the patient’s heart by combining the electrical measurements and the anatomical measurements. The map of the patient’s heart can be stored in a local memory of the surgical systems 110 and transmitted to the local server 120 using the private network 105. In some cases, the map and / or measurements can be transmitted to the local server 120 immediately upon generation.

[0047] In one embodiment, the surgical system 110 enables a surgeon to perform a cardiac ablation procedure. In some cases, the cardiac ablation procedure can utilize contact force technology and irrigation ablation technology. During the cardiac ablation procedure, the surgical system 110 acquires and stores information about the patient's heart and the ablation procedure. For example, the stored information can include information about the arrhythmia of the particular cardiac ablation, the procedure duration, the catheters used, the ablation therapy count and location, the mapping duration, the ablation duration, the ablation power, and other measurable parameters and settings typically collected or optional parameters or settings that can be collected. In addition, the information about the ablation procedure can also include information about the physician performing the procedure, the particular surgical procedure room in which the procedure was performed, and information to identify any support personnel that assisted in the procedure. The surgical system 110 can save the information about the ablation procedure in a local memory and can transmit the information to the local server 120 using the private network 105. In some cases, the information about the ablation procedure is transmitted to the local server 120 during the procedure and in other cases, the information can be transmitted upon completion of the ablation procedure.

[0048] The private network 105 can be any network or system known in the art, such as 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 the surgical system 110 and the local server 120. The private network 105 can be wired, wireless, or a combination thereof. The wired connection can be implemented using Ethernet, Universal Serial Bus (USB), RJ-11, or any other wired connection known in the art. The wireless connection can be implemented using Wi-Fi, WiMAX, and Bluetooth, infrared, cellular networks, satellite, or any other wireless connection method known in the art. Additionally, several networks can work independently or communicate with each other to facilitate communication in the private network 105.

[0049] The local server 120 receives the electrical measurements and anatomical measurements, the mapping diagrams, and the information about the ablation procedure in a local database. In some cases, the local database associates the received data with unique patient identifiable information. In these cases, the local server 120 can save this "individually identifiable health information" in accordance with the Health Insurance Portability and Accountability Act of 1996 (HIPAA).

[0050] Additionally or alternatively, data anonymization or data synthesis can be used. Data anonymization or data synthesis is a process of generating new data, whether based on original real data, patterns of real data, or via the use of random generation to provide synthetic data. Synthetic data can be configured to have greater or lesser analytic utility than the original data set. Synthetic data can also be configured to have greater or lesser privacy, re-identification, or disclosure risk than the original data set. In one embodiment, synthetic data can be based on data of a single patient that has been slightly modified in a random fashion. In another embodiment, synthetic data can be data from a number of patients that are either specifically selected (from a particular data set, a particular physician, a particular hospital, or a particular condition) or randomly selected, that has been averaged to create a synthetic data set. Those skilled in the art will appreciate that there are many ways to generate synthetic data sets within the scope of the present teachings. Generally, there is a tradeoff between analytic utility and privacy risk for any data synthesis technique. Synthetic data can be used where real data is not available or is less than ideal or feasible to use.

[0051] In some cases, the local server 120 is implemented as a physical server. In other cases, the local server 120 is implemented as a virtual server, such as via a public cloud computing provider (e.g., Amazon Web Services (AWS)).

[0052] In some cases, the local server 120 uses machine learning or other artificial intelligence techniques to analyze data stored in the local database. The local server can use machine learning to: 1) consider all previous patients with similar heart conditions and morphology, and associated ablation procedure data and best outcomes from previous patients, in order to recommend a best treatment plan for a heart ablation to be performed; 2) consider the foregoing previous patients and ablation data to modify the treatment plan while performing the heart ablation, and make specific recommendations to the physician during the ablation procedure; and / or 3) evaluate the performance of the surgeon performing the heart ablation according to the treatment plan and the outcomes of the patient. In this way, the physician has the cumulative experience of all previous patients and heart procedures in planning, treating, and evaluating the ablation procedure in order to achieve the best patient outcomes.

[0053] ​In an alternative embodiment, the local server can anonymize the electrical measurements and anatomical measurements, maps, and information about the ablation procedure to form anonymized data. The local server 120 can then transmit the anonymized data to the cloud-based platform 160 via the public network 150 without any risk of non-compliance with HIPPA or GDPR regulations. Using anonymized data greatly expands the database of experiences available to physicians on the public network 150. Those skilled in the art will appreciate that there are many standard techniques for anonymizing data, and a detailed description of such processes is outside the scope of this specification.

[0054] The collection and analysis of data about a particular procedure can include statistical information about the procedure that can be obtained from respective patients, broken down by the particular individual portion of the procedure. This can help physicians and researchers measure the impact of any parameter on any other parameter. In one example, the collection and analysis of such data enables researchers to evaluate any changes introduced into a cardiac catheterization procedure, such as a new catheter, a new version of a catheter, a new drug, a new software, or the like. This can allow researchers to determine whether these changes improve patient outcomes.

[0055] The public network 150 can be any network or system known in the art, including the Internet, 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 the discrete networks 101A-101N and the cloud-based platform 160. The public network 150 can be wired, wireless, or a combination thereof. Wired connections can be implemented using Ethernet, Universal Serial Bus (USB), RJ-11, or any other wired connection known in the art. Wireless connections can be implemented using Wi-Fi, WiMAX, and Bluetooth, infrared, cellular networks, satellite, or any other wireless connection method known in the art. Additionally, several networks can work alone or in communication with one another to facilitate communication in the public network 150.

[0056] ​The cloud-based platform 160 receives anonymized data from each of the local servers 120 of the discrete networks 101A-101N and stores the received information in an anonymized database. In some cases, the local servers 120 use machine learning or other artificial intelligence techniques to analyze the data stored in the local database. The local servers can use machine learning to: 1) consider all previous patients with similar cardiac conditions and morphology along with associated ablation procedure data and best outcomes from previous patients in order to recommend the best treatment protocol for the cardiac ablation to be performed; 2) consider the aforementioned previous patients and ablation data to modify the treatment protocol while performing the cardiac ablation and make specific recommendations to the physician during the ablation procedure; and / or 3) evaluate the performance of the surgeon performing the cardiac ablation according to the treatment protocol and the patient’s outcome. In this way, the physician has the cumulative experience of all previous patients and cardiac procedures in planning, treating, and evaluating the ablation procedure in order to achieve the best patient outcome.

[0057] In many cases, the cloud-based platform 160 also provides an entry point for the third parties 140 to query the data stored in the anonymized database via the public network 150. In some cases, the third parties 140 can use a standard internet browser to access the entry point of the cloud-based platform 160. In other cases, the third parties 140 need a dedicated application to access the entry point of the cloud-based platform 160.

[0058] The cloud-based platform can also provide access to the information in the anonymized database to an augmented reality (AR) / virtual reality (VR) system 130. The AR / VR system 130 enables training of surgeons to perform cardiac ablation procedures based on the data collected about actual cardiac ablations.

[0059] The AR / VR system 130 enables the user to provide training of interest. While a standard display presents the entire view of the system, the AR / VR system 130 enables providing algorithms to areas where more information is desired as will be described. The AR / VR system 130 enables focusing on specific diagnostic elements, such as sending an ultrasound catheter beam to certain areas, and / or qualifying the mapping system to modify the resolution of the map in the area of interest, such as 1 mm to 0.5 mm or even to 0.1 mm point density to enable viewing the area.

[0060] Figure 2A A process 200A is shown that collects data and performs analysis using retrospective analysis to provide better future therapy. At step 210, electrical and anatomical measurements of a patient’s heart are recorded using the surgical system 110 prior to the patient receiving a cardiac ablation treatment. At step 220, available treatment data collected during a cardiac ablation procedure is recorded by the surgical system.

[0061] At step 230, initial data analysis is performed based on the electrical measurements and anatomical measurements recorded during step 210 and the treatment data collected during step 230. The initial data analysis can identify procedure parameters, which includes information that can be used to identify "similar patients" based on clinical parameters and identify treatment method characteristics (e.g., how much ablation energy to deliver within what distance) based on electrical anatomical parameters.

[0062] At step 240, additional information about the patient can be obtained. For example, patient history and patient ongoing care, such as by filling out a form or via a connection with other systems, such as from an EMR, from patient feedback, and from cardiac detection devices, for example, received manually.

[0063] At step 250, additional analysis is performed based on the information collected from step 230, step 220, and step 210 to define "virtual patient" characteristics, treatment methods, and "expected outcomes." At step 260, a similar "virtual patient" group is identified. At step 260, the similar "virtual patient" group is identified by comparing the "virtual patient" defined at step 250 to previously defined virtual patients.

[0064] At step 270, the therapy is scored based on available long-term clinical outcomes, including short, medium, and long-term outcome timeframes. At step 280, a therapy for the similar "virtual patient" group is recommended based on the score calculated at step 270. Additional long-term treatment data about the patient is then collected at step 290. The long-term treatment data can include a starting score, which is based on 1-year ongoing case data, and a re-score thereafter, which is based on 3-year ongoing case data. Then, as this additional data is collected, step 270 uses the additional data to update the therapy score.

[0065] Figure 2B A process 200B for generating a cardiac ablation treatment plan is shown, in accordance with certain embodiments. At step 205, electrical measurements of a patient's heart are obtained by the surgical system 110. At step 220, anatomical measurements of the patient are acquired by the surgical system 110. Based on the electrical measurements and anatomical measurements acquired at steps 205 and 220, the surgical system 110 generates a map of the patient's heart at step 225.

[0066] At step 235, the map of the patient's heart, the anatomical measurements and electrical measurements of the patient's heart are compared to data collected in previous cardiac ablation procedures. Such different types of data are described in detail below. Differences and correlations are determined, and a treatment plan for a cardiac ablation procedure is then generated at step 245 and displayed by the surgical system 110 at step 255.

[0067] Optionally, at step 265, the parameters of the surgical system 110 are automatically configured to perform the cardiac ablation according to the treatment plan generated at step 245. The parameters of the surgical system 110 that can be configured include the intensity, duration, number of times, and location of the ablation. If the ablation is performed by a physician, the physician uses the treatment plan as a guide for the procedure. Alternatively, if the ablation is assisted or performed in whole or in part by a surgical robot, the treatment plan can be a path map for the automated surgical robot system; that is, the location and duration of each ablation is input into the automated surgical robot system and the treatment plan is followed. Such a surgical system is the Monarch surgical robot system by Auris Health, Inc. (Redwood City, CA).

[0068] In some cases, steps 235 and 245 can be performed by the surgical system 110, which transmits the maps of the patient’s heart, the anatomical measurements of the patient’s heart, and the electrical measurements to the local server 120. The local server 120 can utilize machine learning to compare the information received from the surgical system to data stored in the local database. The local server 120 then transmits the results of the comparison back to the surgical system 110 via the private network 105.

[0069] In other cases, steps 235 and 245 can be performed by the surgical system 110, which transmits the maps of the patient’s heart, the anatomical measurements of the patient’s heart, and the electrical measurements to the cloud-based platform 160 via the local server 120. The local server 120 removes any “individually identifiable health information” prior to transmitting the information to the cloud platform 160. The cloud-based platform 160 then utilizes machine learning to compare the information received from the surgical system to data stored in the anonymized database. The cloud-based platform 160 transmits the results of the comparison back to the surgical system 110 via the public network 150 and the local server 120.

[0070] During the procedure, the system can collect a large amount of information about the procedure, which can be broken down and later analyzed by the physician during the evaluation phase.

[0071] Figure 3 A process 300 for performing a cardiac ablation treatment is shown in accordance with certain embodiments. At step 310, a cardiac ablation treatment plan is received by the surgical system 110. In many cases, the cardiac ablation treatment plan is generated using the process 200. At step 320, the surgeon performs the cardiac ablation according to the treatment plan using the surgical system 110. At step 330, the surgical system 110 determines whether additional ablations are needed based on the treatment plan.

[0072] If no additional ablation is needed, the surgical system 110 obtains post-ablation electrical measurements and anatomical measurements of the patient’s heart at step 370. Optionally, the post-ablation electrical measurements and anatomical measurements are displayed by the surgical system 110 at step 370. At step 390, the post-ablation electrical measurements and anatomical measurements, along with information about the ablation procedure, are transmitted to the local server 120 via the private network 105.

[0073] If additional ablation is needed, the surgical system 110 obtains additional electrical measurements and anatomical measurements of the patient’s heart at step 340. The additional electrical measurements and anatomical measurements are then compared to the treatment plan at step 350. Based on the comparison, the treatment plan is revised at step 360. The treatment plan can need to be revised due to errors or inaccuracies in the specific ablation performed at step 320 or changes in the patient’s physiology. Optionally, at step 365, the parameters of the surgical system 110 can be automatically configured to perform cardiac ablation according to the revised treatment plan generated at step 360. The parameters of the surgical system include the intensity, duration, number of times, and location of the ablation. At step 320, the surgeon can perform additional ablation according to the revised treatment plan. In this way, if unexpected difficulties arise during the procedure, the system can make “on the fly” recommendations to the physician for improvements or alternative solutions, allowing the physician multiple treatment options. The system can evaluate each portion of the procedure and provide continuous feedback to the physician. For example, if the treatment plan calls for an ablation at location X and the physician deviates from that plan, even slightly, an ablation plan that modifies for that deviation can be necessary.

[0074] In an alternative embodiment, the system can provide a detailed “grid” on the heart, allowing the physician to be guided to a specific location with high accuracy. An example grid is shown in FIG. 19C, discussed below. Such a grid can allow for easier comparison and analysis of cardiac maps between patients to allow for the suggestion of local activation time (LAT) maps, cycle length (CL) maps, and ripple frequency, and their combination to determine if the ongoing treatment plan is successful. It should be noted that the grid shown as an illustration is greatly simplified for purposes of explanation, and a grid used for actual ablation can have a much greater granularity and thus greater resolution. However, the same concept applies regardless of the granularity of the map.

[0075] Additionally, the system allows for follow-up information to be recorded and taken into account via, for example, the stopper / recorder to obtain long-term data that allows processes and procedures to be associated with positive patient outcomes.

[0076] The system also allows for the inclusion of patient "forms" to collect additional information about the patient that can not be specifically relevant to the patient's cardiac physiology. Form functionality and information can include, for example, a medical history form to determine the patient's age, weight, medications, etc. These forms can allow for the collection and recommendation of short-term (e.g., 12 weeks), medium-term (e.g., 1 year), and / or long-term (e.g., 3-5 years or more) continuity of care information.

[0077] With such data and AI / machine learning algorithms, the system can develop predictive algorithms based on different "categories" of people, diseases, etc. In some cases, steps 340 and 350 are performed by the surgical system 110, transmitting the additional electrical measurements and anatomical measurements to the local server 120. The local server 120 can utilize machine learning to compare the information received from the surgical system to data stored in a local database. The local server 120 then transmits the comparison results back to the surgical system 110 via the private network 105.

[0078] In other cases, steps 340 and 350 are performed by the surgical system 110, transmitting the additional electrical measurements and anatomical measurements to the cloud-based platform 160 via the local server 120. In this example, the local server 120 removes any "individually identifiable health information" prior to transmitting the information to the cloud platform 160. The cloud-based platform 160 utilizes machine learning to compare the information received from the surgical system to data stored in an anonymized database. The cloud-based platform 160 then transmits the comparison results back to the surgical system 110 via the public network 150 and the local server 120.

[0079] Figure 4 A graphical depiction of an artificial intelligence system 400 is shown in accordance with the embodiments disclosed herein. The system 400 includes data 410, a machine 420, a model 430, a plurality of outcomes 440, and underlying hardware 450. The system 400 works in the following manner: the data 410 is used to train the machine 420 while the model 430 is constructed to enable the prediction of the plurality of outcomes 440. The system 400 can work with respect to the hardware 450. In such a configuration, the data 410 can be related to the hardware 450 and can originate from, for example, the surgical system 110. For example, the data 410 can be data being generated or output data associated with the hardware 450. The machine 420 can work as or be associated with a controller or data collector associated with the hardware 450. The model 430 can be configured to model the operation of the hardware 450 and the data 410 collected from the hardware 450 in order to predict outcomes achieved by the hardware 450. Using the predicted outcomes 440, the hardware 450 can be configured to provide a certain desired outcome 440 from the hardware 450.

[0080] Figure 5 It shows in Figure 4 Method 500 is executed in an artificial intelligence system. Method 500 includes collecting data from hardware at step 510. This data may include currently collected data from the hardware, historical data from the hardware, or other data. For example, the data may include measurements taken during a surgical procedure and may be correlated with the outcome of the procedure. For example, the temperature of the heart may be collected and correlated with the outcome of a cardiac procedure.

[0081] At step 520, method 500 includes training the machine on hardware. This training may include analysis and correlation of the data collected in step 510. For example, in the case of the heart, the data on temperature and outcomes may be trained to determine whether there is a correlation or relationship between the heart's temperature and the outcome during the procedure.

[0082] At step 530, method 500 includes building a model on data associated with the hardware. Building the model may include physical hardware or software modeling, algorithmic modeling, etc., as will be described below. This modeling may attempt to represent the collected and trained data.

[0083] At step 540, method 500 includes predicting the outcome of a model associated with the hardware. This prediction of the outcome may be based on a trained model. For example, in the case of the heart, if a temperature between 97.7 and 100.2 during the procedure produces a positive outcome, the outcome can be predicted based on the heart's temperature during the procedure in a given procedure. While this model is basic, it is provided for illustrative purposes and to enhance understanding of the invention.

[0084] The systems and methods of this invention are used to train machines, build models, and predict outcomes using algorithms. These algorithms can be used to solve trained models and predict hardware-related results. These algorithms can generally be categorized into classification, regression, and clustering algorithms.

[0085] For example, classification algorithms are used to categorize a dependent variable (which is the variable being predicted) into multiple classes and predict the class (dependent variable) given an input. Therefore, classification algorithms are used to predict outcomes from a fixed set of predefined results. Classification algorithms can include Naive Bayes, decision trees, random forest classifiers, logistic regression, support vector machines, and k nearest neighbors.

[0086] Generally speaking, the Naive Bayes algorithm follows Bayes' theorem and employs a probabilistic approach. It should be understood that other probability-based algorithms can also be used, and these algorithms typically operate using similar probabilistic principles to those described below for the exemplary Naive Bayes algorithm.

[0087] Figure 6 An example of the probabilities of a Naive Bayes calculation is shown. The probabilistic approach of Bayes' theorem essentially means that the algorithm has a set of prior probabilities for each class of the target, rather than jumping directly into the data. After the input data, the Naive Bayes algorithm can update the prior probabilities to form the posterior probabilities. This is given by the following equation:

[0088]

[0089] This Naive Bayes algorithm, and the Bayes algorithm in general, can be useful when there is a need to predict whether your input belongs to n classes of a given list. The probabilistic approach can be used because the probability of all n classes will be fairly low.

[0090] For example, as shown in FIG. 6, a person plays golf, which depends on factors including the outside weather, as shown in a first data set 610. The first data set 610 shows the weather in a first column and the result of playing golf associated with that weather in a second column. In a frequency table 620, the frequency of certain events occurring is generated. In the frequency table 620, the frequency of a person playing golf or not playing golf under each weather condition is determined. From this, a likelihood table is compiled for generating initial probabilities. For example, the probability of the weather being overcast is 0.29, while the general probability of playing is 0.64. Figure 6 Posterior probabilities can be generated from the likelihood table 630. These posterior probabilities can be configured to answer questions about the weather conditions and whether or not to play golf under those weather conditions. For example, the probability of playing golf outside on a sunny day can be articulated through the Bayes equation:

[0091] P (Yes | Sunny) = P (Sunny | Yes) * P (Yes) / P (Sunny)

[0092] From the likelihood table 630:

[0093] P (Sunny | Yes) = 3 / 9 = 0.33,

[0094] P (Sunny) = 5 / 14 = 0.36,

[0095] P (Yes) = 9 / 14 = 0.64.

[0096] Thus, P (Yes | Sunny) = 0.33 * 0.64 / 0.36 or approximately 0.60 (60%).

[0097]

[0098] ​Generally, a decision tree is a tree structure similar to a flowchart, where each outer node represents a test on an attribute and each branch represents the result of that test. Leaf nodes contain the actual predicted labels. A decision tree starts at the root, where attribute values ​​are compared until a leaf node is reached. Decision trees can be used as classifiers when dealing with high-dimensional data and when little time has been spent on data preparation. Decision trees can take the form of simple decision trees, linear decision trees, algebraic decision trees, deterministic decision trees, stochastic decision trees, nondeterministic decision trees, and quantum decision trees. The following sections... Figure 7 An example decision tree is provided in the document.

[0099] Figure 7 A decision tree for deciding whether to play golf is shown, following the same structure as the Bayesian example above. In the decision tree, the first node 710 checks the weather, thus considering sunny 712, cloudy 714, and rainy 716 as choices to proceed down the decision tree. If the weather is sunny, the tree's leg follows to a second node 720 that checks the temperature. In this example, the temperature at node 720 can be high 722 or normal 724. If the temperature at node 720 is high 722, the prediction is "No" (don't play) 723 for golf. If the temperature at node 720 is normal 724, the prediction is "Yes" (play) 725 for golf.

[0100] Furthermore, starting from the first node 710, if the result is cloudy 714, then it is "yes" (hitting) 715 golf ball.

[0101] Starting with the first node (weather 710), the result of rainy day 716 leads to a third node (730) checking the temperature. If the temperature at the third node 730 is normal 732, then "yes" (play) 733. If the temperature at the third node 730 is low 734, then "no" (don't play) 735.

[0102] According to the decision tree, a golfer will play golf if the weather is cloudy (715), sunny at normal temperature (725), or rainy at normal temperature (733). However, a golfer will not play golf if the weather is sunny and hot (723) or rainy and cold (735).

[0103] A random forest classifier is a committee of decision trees, where each decision tree is fed a subset of attributes from the data and makes predictions based on that subset. The mode of the actual predictions of the decision trees is considered to provide the final random forest answer. Random forest classifiers typically mitigate the overfitting present in independent decision trees, resulting in a more robust and accurate classifier.

[0104] Figure 8An exemplary random forest classifier for classifying the colors of clothing is shown. Figure 8 As shown, the random forest classifier comprises five decision trees 8101, 8102, 8103, 8104, and 8105 (collectively or generally referred to as decision tree 810). Each tree is designed to classify the color of clothing. No discussion of each tree and the decisions made is provided, as each individual tree is typically used as... Figure 7 The decision trees are used for operation. In this example, three of the five trees (8101, 8102, 8104) determine that the clothing is blue, one tree determines that the clothing is green (8103), and the remaining tree determines that the clothing is red (8105). The random forest takes these actual predictions from the five trees and calculates the mode of these actual predictions to provide the random forest answer that the clothing is blue.

[0105] Logistic regression is another algorithm used for binary classification tasks. Logistic regression is based on the logistic function (also known as the sigmoid function). This sigmoid curve can take any real-valued number and maps it between 0 and 1, asymptotically approaching those limits. Logistic models can be used to model the probability of a given category or event, such as pass / fail, win / lose, live / die, or healthy / sick. This can be extended to modeling multiple categories of events, such as determining whether an image contains a cat, dog, lion, etc. Each object detected in the image is assigned a probability between 0 and 1, where the sum of these probabilities is one.

[0106] In a logistic model, the logarithmic odds (logarithm of the odds) of a value labeled "1" is a linear combination of one or more independent variables ("predictors"); these variables can each be binary variables (two categories, encoded by indicator variables) or continuous variables (any real value). The corresponding probability of a value labeled "1" can vary between 0 (definitely the value "0") and 1 (definitely the value "1"), hence the label; the function that converts the logarithmic odds to probabilities is a logistic function, hence this name. The unit of measurement for the logarithmic odds scale is called the sublogarithm, derived from the logistic unit, hence this alternative name. Similar models with different sigmoid functions instead of logistic functions can also be used, such as probabilistic models; the defining characteristic of a logistic model is that the odds of a given outcome are multiplicatively scaled by increasing one of the independent variables at a constant rate, where each independent variable has its own parameter; for a binary dependent variable, this summarizes the odds ratio.

[0107] In a binary logistic regression model, the dependent variable has two levels (on a categorical scale). Outputs with more than two values are modeled by multinomial logistic regression, and if the multiple categories are ordered, by ordinal logistic regression (e.g., proportional odds ordinal logistic model). Logistic regression models themselves simply model the probability of the output given the inputs, and do not perform statistical classification (it is not a classifier), but it can be used to act as a classifier, e.g., by choosing a cutoff value and classifying inputs with probabilities greater than the cutoff value as one class, and inputs with probabilities below the cutoff value as another class; this is the general way of making binary classifiers.

[0108] Figure 9 An example logistic regression is shown. This example logistic regression enables prediction of an outcome based on a set of variables. For example, based on a person's average grade point average, a result of being accepted by a school can be predicted. The past history of the average grade point average and the relationship to being accepted enables the prediction to occur. Figure 9 The logistic regression enables analysis of the average grade point average variable 920 to predict an outcome 910 defined by 0 to 1. At the low end of the S-curve 930, the average grade point average 920 predicts a result 910 of not being accepted. While at the high end of the S-curve 940, the average grade point average 920 predicts a result 910 of being accepted. Logistic regression can be used to predict house values, customer lifetime values in the insurance industry, etc.

[0109] Support vector machines (SVMs) can be used to classify data with a margin between two classes that are as far apart as possible. This is referred to as maximum margin separation. SVMs can consider support vectors when drawing a hyperplane, unlike linear regression, which uses the entire dataset for this purpose.

[0110] Figure 10 An example support vector machine is shown. In the example SVM 1000, data can be classified into two different categories represented as squares 1010 and triangles 1020. The SVM 1000 operates by drawing a random hyperplane 1030. This hyperplane 1030 is monitored by comparing the distance (shown with line 1040) between the hyperplane 1030 and the nearest data points 1050 from each category. The data points 1050 that are nearest to the hyperplane 1030 are referred to as support vectors. The hyperplane 1030 is drawn based on these support vectors 1050, and the best hyperplane has the maximum distance from each of these support vectors 1050. The distance between the hyperplane 1030 and the support vectors 1050 is referred to as the margin.

[0111] The SVM 1000 can be used to classify data by using a hyperplane 1030 such that the distance between the hyperplane 1030 and the support vectors 1050 is maximized. For example, such an SVM 1000 can be used to predict heart disease.

[0112] K-Nearest Neighbors (KNN) refers to a set of algorithms that generally make no assumptions about the underlying data distribution and perform a fairly short training phase. Generally, KNN uses many data points that are divided into several categories to predict the classification of a newly sampled point. Operationally, KNN specifies an integer N that has the new sample. The N entries in the model of the system that are closest to the new sample are selected. The most common classification of these entries is determined and assigned to the new sample. KNN generally requires storage space to increase as the training set increases. This also means that the estimation time increases proportionally to the number of training points.

[0113] In regression algorithms, the output is a continuous quantity, so regression algorithms can be used where the target variable is a continuous variable. Linear regression is a general example of a regression algorithm. Linear regression can be used to estimate true qualities (house cost, number of calls, all buyout transactions, etc.) from one or more consistent variables. A connection between the variables and the results is created by fitting the best line (and thus fitting linear regression). This best fitting line is called the regression line and is expressed by the direct condition Y = a * X + b. Linear regression is best used in methods involving low-dimensional numbers.

[0114] Figure 11 An example linear regression model is shown. In this model, a prediction variable 1110 is modeled with respect to a measured variable 1120. A collection of instances of the prediction variable 1110 and the measured variable 1120 are plotted as data points 1130. The data points 1130 are then fitted with a best fit line 1140. The best fit line 1140 is then used for subsequent predictions given the measured variable 1120, which is used to predict the prediction variable 1110 for that instance. Linear regression can be used to model and predict in financial portfolios, income forecasting, real estate, and traffic in terms of estimating arrival times.

[0115] Clustering algorithms can also be used to model and train a data set. In clustering, inputs are assigned into two or more clusters based on feature similarity. Clustering algorithms generally learn patterns and useful insights from data without any guidance. For example, unsupervised learning algorithms such as K-means clustering can be used to cluster viewers into similar groups based on their interests, age, geography, etc.

[0116] K-means clustering is generally considered a simple unsupervised learning method. In K-means clustering, similar data points can be grouped together and bound in clusters. One method for binding data points together is by computing the centroid of the group of data points. In determining the effective clusters, in K-means clustering, the distance between each point and the centroid of the cluster is evaluated. Based on the distance between the data point and the centroid, the data is assigned to the nearest cluster. The goal of clustering is to determine the intrinsic grouping in a set of unlabeled data. The "K" in K-means represents the number of clusters formed. The number of clusters, which is essentially the number of classes into which new data instances can be classified, can be determined by the user. For example, the determination can be performed using feedback during training and looking at the size of the clusters.

[0117] K-means is primarily used when the dataset has points that are different and well spaced; otherwise, if the clusters are not spaced, the modeling can make the clusters inaccurate. Additionally, K-means can be avoided in cases where the dataset contains a large number of outliers or the dataset is non-linear.

[0118] Figure 12 K-means clustering is illustrated. In K-means clustering, data points are plotted and a value of K is assigned. For example, for K = 2 in Figure 12 the data points are plotted as depicted 1210. The points are then assigned to similar centers at step 1220. Cluster centroids are identified as depicted 1230. Once the centroids are identified, the points are reassigned to clusters to provide a minimum distance between the data points and the respective cluster centroids as depicted 1240. New centroids of the clusters can then be determined as depicted 1250. As the data points are reassigned to the clusters, new cluster centroid formation, iteration, or a series of iterations can occur to minimize the size of the clusters and determine the centroids of the best centroids. Then, when a new data point is measured, the new data point can be compared to the centroids and clusters to identify with the cluster.

[0119] Ensemble learning algorithms can be used. These algorithms use multiple learning algorithms to obtain better predictive performance than can be obtained from any of the constituent learning algorithms alone. Ensemble algorithms perform the task of searching through the hypothesis space to find a good hypothesis that will make good predictions for a particular problem. Finding a good hypothesis can be very difficult even if the hypothesis space contains a hypothesis that is very well suited for the particular problem. Ensemble algorithms combine multiple hypotheses to form a better hypothesis. The term ensemble is generally reserved for methods that use the same base learner to generate multiple hypotheses. A broader term for multiple classifier systems also encompasses mixtures of hypotheses that are not induced by the same base learner.

[0120] Evaluating the predictions of an ensemble generally requires more computation than evaluating the predictions of an individual model, so ensembles can be thought of as a way to compensate for a poor learning algorithm by performing a large amount of extra computation. Fast algorithms such as decision trees are often used in ensemble methods, e.g., random forests, although slower algorithms can also benefit from ensemble techniques.

[0121] The ensemble itself is a supervised learning algorithm, as the ensemble can be trained and then used to make predictions. Thus, a trained ensemble represents a single hypothesis. However, this hypothesis is not necessarily contained within the hypothesis space of the models that built the hypothesis. Thus, it can be shown that ensembles have greater flexibility in what they can represent. In theory, this flexibility can enable these ensembles to fit the training data better than individual models, but in practice, some ensemble techniques, especially bagging, tend to reduce problems related to overfitting the training data.

[0122] Empirically, ensemble algorithms tend to produce better results when there is significant diversity among the models. Thus, many ensemble methods attempt to promote diversity among the models that they combine. While it is non-intuitive, it is possible to use more random algorithms (like random decision trees) to produce stronger ensembles than very deliberate algorithms (like entropy-reducing decision trees). However, using a variety of powerful learning algorithms has been shown to be more effective than using techniques that attempt to discard models to promote diversity.

[0123] The number of component classifiers of an ensemble has a large impact on the accuracy of the predictions. Determining the ensemble size and volume of a large data stream in advance and the speed makes this even more important for online ensemble classifiers. Theoretical frameworks suggest that there is an ideal number of component classifiers for an ensemble, such that having more or fewer classifiers will decrease accuracy. Theoretical frameworks suggest that using an independent component classifier of the same number as the class labels gives the highest accuracy.

[0124] Some common types of ensembles include Bayesian optimal classifier, bootstrap aggregation (bagging), boosting, Bayesian model averaging, Bayesian model combination, model storage, and stacking. Figure 13 An example ensemble learning algorithm is shown, where bagging is performed in parallel 1310 and boosting is performed sequentially 1320.

[0125] A neural network is a network or circuit of neurons, or in modern terms, an artificial neural network composed of artificial neurons or nodes. The connections of biological neurons are modeled as weights. Positive weights reflect excitatory connections, while negative values represent inhibitory connections. Inputs are modified by the weights and summed using a linear combination. An activation function can control the amplitude of the output. For example, the range of acceptable outputs is often between 0 and 1, or the range can be between -1 and 1.

[0126] These artificial networks can be used for predictive modeling, adaptive control and applications, and can be trained via data sets. Self-learning from experience can occur within the network, which can draw conclusions from complex and seemingly unrelated groups of information.

[0127] For completeness, a biological neural network consists of one or more groups of chemically connected or functionally related neurons. Individual neurons can be connected to many other neurons, and the total number of neurons and connections in a network can be extensive. Connections, called synapses, are usually formed by axon to dendrite, but dendritic synapses and other connections are also possible. In addition to electrical signals, there are other forms of signals caused by neurotransmitter diffusion.

[0128] Artificial intelligence, cognitive modeling, and neural networks are information processing paradigms induced by the way biological nervous systems process data. Artificial intelligence and cognitive modeling attempt to mimic some of the properties of biological neural networks. In the field of artificial intelligence, artificial neural networks have been successfully applied to speech recognition, image analysis, and adaptive control in order to build software agents or autonomous robots (in computers and video games).

[0129] A neural network (NN) is a set of interconnected natural or artificial neurons, referred to as artificial neural networks (ANN) or simulated neural networks (SNN), that use mathematical or computational models based on computational connection methods for information processing. In most cases, ANNs are adaptive systems that change their structure based on external or internal information flowing through the network. The more practical term neural network is a nonlinear statistical data modeling or decision-making tool. These terms neural network can be used to model complex relationships between inputs and outputs or find patterns in data.

[0130] Artificial neural networks involve networks of simple processing elements (artificial neurons) that can exhibit complex global behavior determined by the connections between the processing elements and the element parameters.

[0131] A classical type of artificial neural network is the recurrent Hopfield network. The utility of artificial neural network models lies in their ability to infer functions from observations and also to use the inferred functions. Unsupervised neural networks can also be used to learn a representation of the inputs that captures the salient features of the input distribution, and recent deep learning algorithms can implicitly learn a distribution function of the observed data. Learning in neural networks is particularly useful in applications where the complexity of the data or task makes it impractical to manually design such functions.

[0132] Neural networks can be used in different domains. Tasks to which artificial neural networks are applied often fall within the following broad categories: function approximation or regression analysis, including time series prediction and modeling; classification, including pattern and sequence recognition, novelty detection, and sequential decision making; and data processing, including filtering, clustering, blind signal separation, and compression.

[0133] Application areas for ANNs include nonlinear system identification and control (vehicle control, process control), game play and decision making (chess, chat, poker), pattern recognition (radar systems, face recognition, object recognition), sequence recognition (gesture, speech, handwriting text recognition), medical diagnosis, financial applications, data mining (or knowledge discovery in databases, "KDD"), visualization, and email spam filtering. For example, a semantic feature map of user interests can be created from pictures trained for object recognition.

[0134] Figure 14 An example neural network is shown. In this neural network, there is an input layer represented by a number of inputs such as 14101 and 14102. The inputs 14101, 14102 are provided to a hidden layer depicted as including nodes 14201, 14202, 14203, 14204. These nodes 14201, 14202, 14203, 14204 are combined to produce an output 1430 in an output layer. The neural network performs simple processing via a hidden layer of simple processing elements (nodes 14201, 14202, 14203, 14204) that can represent a complex global behavior determined by connections between the processing elements and element parameters.

[0135] Figure 14 Neural networks can be implemented in hardware. As shown in Figure 15 A hardware-based neural network is shown.

[0136] Cardiac arrhythmias and, in particular, atrial fibrillation, are consistently common and dangerous medical conditions, particularly in the elderly. For a patient with normal sinus rhythm, the heart, which is composed of atrial, ventricular, and excitatory conduction tissue, beats in a synchronized, patterned manner under the influence of electrical stimulation. For a patient with a cardiac arrhythmia, abnormal regions of cardiac tissue do not follow the synchronized beating cycle associated with normal conduction tissue as do patients with normal sinus rhythm. Instead, the abnormal regions of cardiac tissue abnormally conduct to adjacent tissue, disrupting the cardiac cycle into a non-synchronized rhythm. Such abnormal conduction has been previously known to occur at various regions of the heart, such as in the sinoatrial (SA) node region, along the conduction pathway of the atrioventricular (AV) node and bundle of His, or in the myocardial tissue forming the walls of the ventricular and atrial chambers.

[0137] Arrhythmias, including atrial arrhythmias, can be of the multiple wavelet reentry type, characterized by multiple asynchronous loops of electrical impulses that are dispersed around the atrial chambers and that generally self-propagate. Alternatively, or in addition to the multiple wavelet reentry type, arrhythmias can have a focal origin, such as when an isolated region of tissue within the atrium autonomously beats in a rapidly repeating manner. Ventricular tachycardia (V-tach or VT) is a tachycardia or rapid heart rhythm originating from one of the ventricles. It is a potentially life-threatening arrhythmia because it can lead to ventricular fibrillation and sudden death.

[0138] Atrial fibrillation is a type of arrhythmia that occurs when the normal electrical impulses produced by the sinoatrial node are overwhelmed by disorganized electrical impulses originating in the atria and pulmonary veins, which cause irregular impulses to be transmitted to the ventricles. Irregular heartbeats result, and can last from a few minutes to weeks, or even years. Atrial fibrillation (AF) is usually a chronic condition that slightly increases the risk of death, usually from stroke. The risk increases with age. About 8% of people over 80 years of age have some degree of AF. Atrial fibrillation is usually asymptomatic and not life-threatening by itself, but it can cause palpitations, weakness, dizziness, chest pain, and congestive heart failure. The risk of stroke is increased during AF because blood can pool and form clots in the poorly contracting atria and left atrial appendage. First-line treatment for AF is drug therapy that can slow the heart rate or restore normal heart rhythm. In addition, people with AF are usually given an anticoagulant to prevent their risk of stroke. The use of such anticoagulants is accompanied by its own risk of internal bleeding. For some patients, drug therapy is not sufficient, and their AF is considered drug refractory, i.e., intractable to standard drug intervention. Synchronized electrical cardioversion can also be used to restore AF to normal rhythm. Alternatively, AF patients can be treated by catheter ablation.

[0139] Catheter ablation-based treatment can include mapping electrical properties of cardiac tissue, particularly endocardial and cardiac volume, and selectively ablating cardiac tissue by applying energy. Cardiac mapping, such as creating a map of electrical potentials (voltage map) or a map of time of arrival (local time activation (LAT) map) of wave propagation along cardiac tissue to various tissue locations, can be used to detect local cardiac tissue dysfunction ablation, such as those based on cardiac mapping, can stop or modify unwanted electrical signals from propagating from one part of the heart to another.

[0140] Ablation procedures damage unwanted electrical pathways by creating non-conducting ablation lesions. Various forms of energy delivery for creating ablation lesions have been disclosed, including the use of microwave, laser, and more commonly, radiofrequency energy to create conduction blocks along the walls of cardiac tissue. In a two-step procedure of mapping then ablation, electrical activity at various points in the heart is sensed and measured, typically by inserting a catheter containing one or more electrical sensors (or electrodes) into the heart and taking data at multiple points. The data is then used to select endocardial target regions for ablation.

[0141] As clinicians treat increasingly challenging conditions such as atrial fibrillation and ventricular tachycardia, cardiac ablation and other cardiac electrophysiology procedures have become increasingly complex. Treatment of complex arrhythmias currently relies on the use of three-dimensional (3D) mapping systems in order to reconstruct the anatomical structure of the ventricle of interest.

[0142] For example, cardiologists rely on software, such as the CARTO® 3D mapping system by Biosense Webster, Inc. (Diamond Bar, Calif.) to analyze intracardiac EGM signals and determine ablation points for treating a wide range of cardiac conditions, including atypical atrial flutter and ventricular tachycardia. The 3D mapping system's complex complex fractionated atrial electrogram (CFAE) module analyzes intracardiac EGM signals and determines ablation points for treating a wide range of cardiac conditions, including atypical atrial flutter and ventricular tachycardia.

[0143] The 3D map can provide multiple pieces of information about the electrophysiological properties of the tissue, which represent the anatomical and functional substrate of these challenging arrhythmias.

[0144] Cardiomyopathies with different etiologies (hypoxia, dilated (DCM), hypertrophic cardiomyopathy (HCM), arrhythmogenic right ventricular dysplasia (ARVD), left ventricular noncompaction (LVNC), etc.) have an identifiable substrate characterized by regions of unhealthy tissue surrounded by functionally normal cardiomyocytes.

[0145] Figure 16is a diagram of an exemplary system 1620 that can implement one or more features of the subject disclosure. All or a portion of the system 1620 can be used to collect information for a training data set, and / or all or a portion of the system 1620 can be used to implement a trained model. The system 1620 can include a component configured to damage a tissue region of an in vivo organ, such as a catheter 1640. The catheter 1640 can also be further configured to obtain biometric data. While the catheter 1640 is shown as a sharp catheter, it should be understood that a catheter of any shape including one or more elements (e.g., electrodes) can be used to implement the embodiments disclosed herein. The system 1620 includes a probe 1621 having a shaft that can be navigated by a physician 1630 into a body part of a patient 1628, such as a heart 1626, lying on a table 1629. According to embodiments, multiple probes can be provided, however, for the sake of brevity, a single probe 1621 is described herein, but it should be understood that the probe 1621 can be representative of multiple probes. As shown, the physician 1630 can insert the shaft 1622 through a sheath 1623 while manipulating the distal end of the shaft 1622 using a manipulator 1632 near the proximal end of the catheter 1640 and / or deflecting from the sheath 1623. As shown in the inset 1625, the catheter 1640 can fit over the distal end of the shaft 1622. The catheter 1640 can be inserted through the sheath 1623 in a collapsed state and then can be deployed within the heart 1626. As disclosed further herein, the catheter 1640 can include at least one ablation electrode 1647 and a catheter needle 1648. Figure 16

[0146] According to exemplary embodiments, the catheter 1640 can be configured to ablate a tissue region of a chamber of the heart 1626. The inset 1645 shows the catheter 1640 within a chamber of the heart 1626 in a magnified view. As shown, the catheter 1640 can include at least one ablation electrode 1647 coupled to a main body of the catheter. According to other exemplary embodiments, multiple elements can be coupled via a long strip that forms a shape of the catheter 1640. One or more other elements (not shown) can be provided that can be any element configured to ablate or obtain biometric data, and can be an electrode, a transducer, or one or more other elements.

[0147] According to embodiments disclosed herein, an ablation electrode, such as the electrode 1647, can be configured to provide energy to a tissue region of an in vivo organ, such as the heart 1626. The energy can be thermal energy and can cause damage to the tissue region starting from a surface of the tissue region and extending into a thickness of the tissue region.

[0148] ​According to the exemplary embodiments disclosed herein, biometric data may include one or more of the following: LAT, electrical activity, topology, bipolar mapping, dominant frequency, impedance, etc. Local excitation time can be a time point of threshold activity corresponding to local excitation, calculated based on a normalized initial start point. Electrical activity can be any applicable electrical signal that can be measured based on one or more thresholds and can be sensed and / or amplified based on signal-to-noise ratio and / or other filters. Topology can correspond to the physical structure of a body part or a portion of a body part, and can correspond to variations in the physical structure relative to different parts of the body part or relative to different body parts. Dominant frequency can be a frequency or frequency range that is prevalent in a part of a body part and can differ in different parts of the same body part. For example, the dominant frequency of the pulmonary veins of the heart can differ from the dominant frequency of the right atrium of the same heart. Impedance can be a resistance measurement at a given region of a body part.

[0149] like Figure 16 As shown, probe 1621 and catheter 1640 can be connected to console 1624. Console 1624 may include processor 1641 (such as a general-purpose computer) having suitable front-end and interface circuitry 1638 for transmitting and receiving signals to and from the catheter, as well as other components for controlling system 1620. In some embodiments, processor 1641 may be further configured to receive biometric data, such as electrical activity, and determine whether a given tissue region is conductive. According to one embodiment, the processor may be external to console 1624 and may be located, for example, in a catheter, an external device, a mobile device, a cloud-based device, or may be a stand-alone processor.

[0150] Processor 1641 may include a general-purpose computer that can be programmed with software to perform the functions described herein. The software may be downloaded to the general-purpose computer electronically, for example, via a network, or alternatively or additionally set and / or stored on a non-transitory tangible medium, such as magnetic storage, optical storage, or electronic storage. Figure 16 The exemplary configuration shown can be modified to implement the embodiments disclosed herein. The disclosed embodiments can be applied similarly using other system components and settings. Additionally, system 1620 may include additional components such as elements for sensing electrical activity, wired or wireless connectors, processing and display devices, etc.

[0151] According to one embodiment, the display connected to the processor (e.g., processor 1641) can be located at a remote location such as a separate hospital or in a separate network of healthcare providers. Further, the system 1620 can be part of a surgical system configured to obtain anatomical measurements and electrical measurements of a patient’s organ, such as a heart, and perform a cardiac ablation procedure. An example of such a surgical system is the NAVIGATOR® system sold by Biosense Webster .

[0152] The system 1620 can also and optionally use ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), or other medical imaging techniques known in the art to obtain biometric data, such as anatomical measurements of a patient’s heart. The system 1620 can use a catheter, electrocardiogram (EKG), or other sensor that measures electrical properties of the heart to obtain electrical measurements. As Figure 16 shown, the biometric data, including the anatomical measurements and electrical measurements, can then be stored in the memory 1642 of the mapping system 1620. The biometric data can be transmitted from the memory 1642 to the processor 1641. Alternatively or additionally, the biometric data can be transmitted to a server 1660, which can be local or remote, using the network 1662.

[0153] The network 1662 can be any network or system known in the art, such as 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 the mapping system 1620 and the server 1660. The network 1662 can be wired, wireless, or a combination thereof. Wired connections can be implemented using Ethernet, Universal Serial Bus (USB), RJ-11, or any other wired connection known in the art. Wireless connections can be implemented using Wi-Fi, WiMAX, and Bluetooth, infrared, cellular networks, satellite, or any other wireless connection method known in the art. Additionally, several networks can work alone or in communication with one another to facilitate communication in the network 1662.

[0154] In some cases, the server 1662 can be implemented as a physical server. In other cases, the server 1662 can be implemented as a virtual server of a public cloud computing provider (e.g., Microsoft Azure, Amazon Web Services ).

[0155] The console 1624 can be connected to body surface electrodes 1643 by cables 1639, which can include adhesive skin patches attached to the patient 1628. The processor, in combination with a current tracking module, can determine position coordinates of the catheter 1640 within a body site of the patient, such as the heart 1626. The position coordinates can be based on impedance or electromagnetic fields measured between the body surface electrodes 1643 and the electrodes 1648 or other electromagnetic components of the catheter 40. Additionally or alternatively, a positioning mat can be located on the surface of the bed 1629 and can be separate from the bed 1629.

[0156] The processor 1641 can 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 EMG signal conversion integrated circuit. The processor 1641 can pass the signals from the A / D ECG or EMG circuit to another processor and / or can be programmed to perform one or more functions disclosed herein.

[0157] The console 1624 can also include an input / output (I / O) communication interface that enables the console to pass signals from and / or to the electrodes 47.

[0158] During the procedure, the processor 1641 can facilitate the presentation of a body site rendering 1635 to the physician 1630 on the display 1627 and store data representative of the body site rendering 1635 in the memory 1642. The memory 1642 can include any suitable volatile and / or non-volatile memory, such as random access memory or a hard drive. In some embodiments, the medical professional 1630 can be able to manipulate the body site rendering 1635 using one or more input devices, such as a trackpad, mouse, keyboard, gesture recognition device, etc. For example, the input device can be used to change the position of the catheter 1640 such that the rendering 1635 is updated. In alternative embodiments, the display 1627 can include a touchscreen that can be configured to accept input from the medical professional 1630 in addition to presenting the body site rendering 1635.

[0159] During a real or virtual procedure, the present system can collect all data related to the procedure. This can include all locations of one or more devices, all measurements, and all treatments provided. In addition, the system can collect any analysis performed prior to or during the procedure and any measurements after the procedure. This can include analysis and any internal analysis performed presented to the operator. By way of example, the data can include catheter location, all recorded ECGs, impedance measurements, contact information, ultrasound images, X-ray images, information about energy delivered during the procedure (including information such as power, temperature generated, and impedance before, during, and after application). Analysis can include all maps, stability parameters, presented images, and calculations. For example, biometric data can include one or more of LAT, electrical activity, topology, bipolar map, dominant frequency, impedance, etc.

[0160] Each procedure or a series of procedures, including all recorded procedures, can be analyzed. Analysis can include checking parameters of one or more procedures to determine how the procedure can be improved. This can include applying any of the methods to Figure 4 Figure 15 By way of example, in calculating improvements, steps similar to a cyclist who records heart rate and pedal rate during a ride and then analyzes that data to determine where force can be increased or decreased to optimize power and improve the outcome of the ride can be performed.

[0161] In one particular embodiment, WACA generation is performed in phases of many ablation points, testing the isolation of the dome, and closing the gap, with a procedure designed to improve wide area circumferential ablation (WACA). CARTO presents the collected ablation points and ablation index values. In performing the analysis, the following steps are performed: identifying points in the first pass, presenting added points in a different color, and presenting details of previously collected points, determining why those points were not good enough, e.g., the catheter was unstable and / or the force was not constant for that period of time. This analysis can help the physician with the ablation technique when the physician needs to focus on the treatment of the patient during the procedure.

[0162] In another particular embodiment for improving the finding of conductive gaps in ablation, channels in scars, the system records ECG data including voltage and LAT, and by analyzing the maps, activation velocity, as in CARTO. Using a system that is detached from the procedure, a simulation can be started that mimics what would happen if a little bit earlier in each location was started to teach the technique about making the analysis where it is important to make the ablation to generate a better ablation technique.

[0163] In the system, all data is good, and via Figure 4 to Figure 15 ​Analysis is necessary to determine what data is useful. As much data as possible can be recorded and after the procedure an analysis algorithm can be developed via the above techniques allowing the data available to be understood. After the algorithm is defined, aspects of the data can be focused on allowing a subset of data to be collected next.

[0164] Figure 17 A process 1700 for ablating cardiac ablation treatments is shown in accordance with certain embodiments. At step 1710, electrical measurements and anatomical measurements, maps, and information about a specific ablation procedure are retrieved. In some cases, the information is retrieved from the cloud-based platform 160 and in other cases, the information is retrieved from the local server 120 and in further cases, some information can be retrieved from either or both of the cloud-based platform 160 and the local server 120.

[0165] At step 1720, post-ablation electrical measurements and anatomical measurements for the specific ablation procedure are retrieved. At step 1730, a treatment score is generated by comparing the post-ablation electrical measurements and anatomical measurements, electrical measurements and anatomical measurements, maps, and information about the specific ablation procedure to previously performed procedures. In some cases, the comparison is performed using machine learning of the local server 120 based on data stored in a local database. In other cases, the comparison is performed using machine learning of the cloud-based platform based on anonymized data stored in an anonymized database. At step 1740, the treatment score computed in 1740 is displayed.

[0166] Figure 18 A process 1800 for training a surgeon to perform cardiac ablation using an AR / VR system is shown in accordance with certain embodiments. At step 1810, a trainee (e.g., a surgeon trainee) inputs parameters about a heart to start a practice. Exemplary parameters can include a specific cardiac malformation or age and gender of the patient. These parameters can come from both a patient history including imaging (CT / MR) data, stopper data, and from early stages of a procedure including electroanatomical data.

[0167] At step 1820, the anonymized database of cloud-based platform 160 is queried for cardiac ablations matching the parameters, and a particular cardiac ablation is selected based on the query. At step 1830, the electrical measurements and anatomical measurements of the selected ablation are loaded into the AR / VR system. After the measurements are loaded into the AR / VR system, the trainee performs a cardiac ablation according to a virtual ablation protocol at step 1840. In some cases, the trainee can be prompted to select among multiple virtual ablation protocols. At step 1850, the AR / VR system compares the trainee’s performance to the results of a surgeon performing the selected cardiac ablation. At step 1860, a score is calculated and displayed by the AR / VR system based on the comparison in 1850.

[0168] Figure 19A An example of a map 1900A that can be generated by the surgical system 110 is shown. As previously described, the map 1900A is formed from a detailed anatomical map and an electrical map of the patient’s heart. By comparing the map of the current patient to a library of maps of other patients, the system can determine similarities and differences between the map of the current patient and the maps of other patients in the library. The system can then determine the maps in the patient library that provide or lead to the best treatment outcomes for conditions similar to the condition of the current patient. Based on this information, the system can use the determined stored maps in order to calculate the best treatment protocol for the current patient. In a similar manner, when a treatment protocol is performed on a current patient, the system records the ablation treatment and adds this information to the patient library. After the current patient outcome is determined, the outcome is added to the library.

[0169] Figure 19A A standard CARTO® map 1900A is represented. The large number of dots shown in the image represent the locations where points were acquired during the procedure, scars in the areas where material was ablated (darker dots), and the progress of the operation represented by other shading.

[0170] Figure 19B An example of a map 1900B that can be generated by the surgical system 110 is shown. As previously described, the map 1900B is formed from a detailed anatomical map and an electrical map of the patient’s heart. The map 1900B shows additional detail and a higher resolution view of the ablation points from the procedure. An ablation index is shown, and locations related to two pulmonary veins (LIPV and LSPV) are provided. The left atrial appendage (LAA) is also shown. These details provide the relationship of the ablation to the anatomical structures.

[0171] Figure 19CAn illustration of another example of a map 1900B that can be generated by the surgical system 110 is shown. As previously discussed, the map 1900C is formed from a detailed anatomic map and an electrical map of a patient's heart. By comparing the current patient's map to a library of maps of other patients, the system can determine similarities and differences between the current patient's map and other patient's maps present in the library.

[0172] The map 1900C includes a horizontal-vertical grid 1910 overlaid on an image of the patient's heart 1905. In some cases, the size of the cells in the horizontal-vertical grid 1910 is fixed (e.g., 1 mm x 1 mm). In other cases, the horizontal-vertical grid 1910 contains a predetermined number of cells, and the size of the individual cells is scaled based on the size of the patient's heart.

[0173] As discussed herein, the present system 10 provides the ability to collect retrospective anonymized data from multiple hospitals to generate an AI algorithm designed to identify potential gaps in WACA. The system 10 provides continuous improvement in the delivery of WACA by training. The present system 10 provides physicians with detailed analysis of the data so that some of the physicians know how (and where) to improve. Improvements can include improving WACA ablation quality, including producing better ablation in the first pass. The system 10 can result in a reduced probability of case redo and reduce the need for closure gaps after adenosine stress testing.

[0174] The system receives data in the appropriate context, such as a pair of patients with paroxysmal atrial fibrillation in a first case plus, for example, a secondary procedure. The analysis can be reviewed and the physician is able to visualize aspects of the procedure (e.g., stability) that resulted in less than ideal ablation, such as through AR / VR. In this way, the physician can understand how to improve by improving stability in the example case. The system can support the labeling of data, such as by performing an initial labeling. The physician or other worker can be required to correct and approve when modifying the labeling.

[0175] The data collected can be used to test and train physicians and algorithms as described in the example algorithm above. This training enables the identification of locations where ablation improvement can benefit the improvement of WACA continuity. The system 10 is able to analyze the data and provide access to tools that provide a visual analysis of the ablation. The data can also be reviewed by other physicians and experts to assess the results and label the results to improve or correct the labeling, providing improved understanding of the data.

[0176] In the example WACA analysis tool, there is Figure 20A a display 2000A of the RF index Shure point shown in FIG. 20B. Figure 20AThe display 2000A shown in FIG. 20A enables the physician to view the values presented in the CARTO procedure and confirm that these ablation points are for WACA. These points represent the ablations performed and the display shows the RF index of the ablation, i.e., an estimate of the amount of RF energy that has been delivered to the tissue based on the amount of energy delivered, the contact force, and the amount of time the energy was delivered.

[0177] Figure 20B The display 2000B is shown identifying ablation points based on time. The display 2000B enables the physician to identify ablations for gap closure after testing by adenosine stimulation. The physician can view the display 2000B to view the two WACA loop times, i.e., the first path and the correction path (ablation points 2020 and 2022). The physician can mark and / or confirm each WACA loop and correction point. In the case where a secondary ablation is needed or desired, the physician can be able to perform the secondary ablation points on top of the first atrial fibrillation case (not presented on the image).

[0178] In the display 2000B, the points represent the time at which the ablation was formed during the procedure, with the first ablation time designated as zero and times thereafter in units of minutes. The display 2000B enables the physician to understand the order of the ablation points during the ablation and when the procedure was stopped to perform testing. For example, the timing of the loop test, and when the procedure corrected the loop. In the example display 2000B, the ablation points 2020 and the associated ablation points represent the correction of the loop.

[0179] Figure 20C The display 2000C is shown, which shows a stability view of the RF index detail representation. The points are located in space and identify the amount of points near the point of instability marker. The physician can be able to explore the stability of the ablation in each point and understand if there is a correlation between the stability and the location of the correction ablation needed to close the gap.

[0180] The display 2000C represents ablation delivered based on microspots (per mm) and non-ablation points (per partial radius up to 2-5 mm). The display 2000C shows that the catheter is unstable. A stable catheter is identified by the accumulation of the amount of energy delivered in the points to include an RF index in the range of 600 to 700.

[0181] Figure 20D and Figure 20EAdditional displays 2000D and 2000E are shown providing a display of additional parameters to the using physician. These additional parameters include impedance drop, baseline impedance, maximum temperature, etc. In addition, the physician can switch the display to filters by each parameter to improve the analysis. The displays 2000D and 2000E present the energy delivered by the RF generator in Watts. The physician can view the displays 2000D and 2000E according to the protocol to understand the energy in both the posterior wall and the internal wall.

[0182] Figure 21A A graphical representation 2100A of a cardiac ablation treatment plan 2110 is shown. The cardiac ablation treatment plan 2110 identifies a plurality of ablations to be performed on a patient's heart 1905. For example, the cardiac ablation treatment plan 2110 is shown to include four ablations, identified as a first ablation 2101, a second ablation 2102, a third ablation 2103, and a fourth ablation 2104. While four ablations are shown, one of ordinary skill in the art will appreciate that any number of ablations is possible. Additionally, the identification of the ablations as a first, second, third, and fourth ablation does not necessarily indicate the order in which the ablations are performed, but is intended to delineate the different ablations for ease of understanding.

[0183] The cardiac ablation treatment plan 2110 includes the first ablation 2101 to be performed in cell Bl of the horizontal-vertical grid 1910. In addition to the location, the cardiac ablation treatment plan 2110 can identify the type, duration, and intensity of the first ablation 2101. The cardiac ablation treatment plan 2110 also includes the second ablation 2102 to be performed in cell E2 of the horizontal-vertical grid 1910. The type, duration, and intensity of the second ablation 2102 can also be specified in the cardiac ablation treatment plan 2110. The cardiac ablation treatment plan 2110 is shown to include the third ablation 2103 to be performed in cell D4 of the horizontal-vertical grid 1910 and the fourth ablation 2104 to be performed in cell B4 of the horizontal-vertical grid 1910. The type, duration, and intensity of the third ablation 2103 and the fourth ablation 2104 can also be specified in the cardiac ablation treatment plan 2110.

[0184] Figure 21B A graphical representation 2100B of the location of the first ablation performed according to the cardiac ablation treatment plan 2110 is shown. According to the cardiac ablation treatment plan 2110, the first ablation 2101 is to be performed by the surgeon. However, the surgeon in fact performed an ablation 2111. The ablation 2111 includes the location where the ablation was actually performed by the surgeon. The ablation 2111 can also include information regarding the type and intensity of the ablation actually delivered to the patient's heart 1905.

[0185] Figure 21CA graphical representation 2100C of the first revised cardiac ablation treatment protocol 2120 is shown. The first revised cardiac ablation treatment protocol 2120 is recalculated based on the actual ablations 2111. The revised cardiac ablation treatment protocol 2120 includes a revised second ablation 2122 compared to the second ablation 2120 in protocol 2110, a revised third ablation 2123 compared to the third ablation 2103 in protocol 2110, and a revised fourth ablation 2124 compared to the fourth ablation 2102 in protocol 2110. The revised second ablation 2122, the revised third ablation 2123, and the revised fourth ablation 2124 can each include the location, type, duration, and intensity of the respective ablation, respectively. In some cases, the first revised cardiac ablation treatment protocol 2120 can include a different number of ablations than the cardiac ablation treatment protocol 2110.

[0186] Figure 21D A graphical representation 2100D of the location of the second ablation performed according to the first revised cardiac ablation treatment protocol 2120 is shown. According to the first revised cardiac ablation treatment protocol 2120, the surgeon is supposed to perform a second ablation 2122. However, the surgeon in fact performed an ablation 2132. The ablation 2132 includes the location of the ablation actually performed by the surgeon. The ablation 2132 can also include information about the type and intensity of the ablation actually delivered to the patient's heart 2105.

[0187] Figure 21E A graphical representation 2100E of the second revised cardiac ablation treatment protocol 2130 is shown. The second revised cardiac ablation treatment protocol 2130 is recalculated based on the actual ablations 2111 and the actual ablation 2132. The second revised cardiac ablation treatment protocol 2130 includes a further revised third ablation 2133 compared to the third ablation 2123 in protocol 2120, and a further revised fourth ablation 2134 compared to the fourth ablation 2124 in protocol 2120. The further revised third ablation 2133 and the further revised fourth ablation 2144 can each include the location, type, duration, and intensity of the respective ablation. In some cases, the second revised cardiac ablation treatment protocol 2130 can include a different number of ablations than the first revised cardiac ablation treatment protocol 2120.

[0188] Figure 21F A graphical representation 2100F of the location of the third ablation performed according to the second revised cardiac ablation treatment protocol 2130 is shown. According to the revised cardiac ablation treatment protocol 2130, the surgeon is supposed to perform a third ablation according to 2133. However, the surgeon in fact performed an ablation 2143. The ablation 2143 includes the location of the ablation actually performed by the surgeon. The ablation 2143 can also include information about the type and intensity of the ablation actually delivered to the patient's heart 1905.

[0189] Figure 21G A graphical representation 2100G of a third revised cardiac ablation treatment protocol 2140 is shown. The third revised cardiac ablation treatment protocol 2140 is recalculated based on actual ablation 2111, actual ablation 2132, and actual ablation 2143. The third revised cardiac ablation treatment protocol 2140 includes another revised fourth ablation 2144 compared to the fourth ablation 2134 of the protocol 2130. The further revised fourth ablation 2144 can include the location, type, duration, and intensity of the ablation. In some cases, the third revised cardiac ablation treatment protocol 2140 can include a different number of ablations than the second revised cardiac ablation treatment protocol 2130.

[0190] Figure 21H A graphical representation 2100H of the location of a fourth ablation performed according to the third revised cardiac ablation treatment protocol 2140 is shown. According to the revised cardiac ablation treatment protocol 2140, the surgeon should perform a third ablation according to 2144. However, the surgeon in fact performed an ablation 2154. The ablation 2154 includes the location where the surgeon actually performed the ablation. The ablation 2154 can also include information about the type and intensity of the ablation that was actually delivered to the patient’s heart 1905.

[0191] Any of the functions and methods described herein can be implemented in a general purpose computer, a processor, or a processor core. By way of example, suitable processors include: a general purpose processor, a special purpose processor, a conventional processor, a digital signal processor (DSP), a plurality of microprocessors, one or more microprocessors in association with a DSP core, a controller, a microcontroller, Application Specific Integrated Circuits (ASICs), Field Programmable Gate Array (FPGAs) circuits, any other type of integrated circuit (IC), and / or a state machine. Such processors can be made of a hardware description language (HDL) instruction and other intermediate data (such instructions being stored on a computer readable medium) including network tables, which can be used to configure a manufacturing process to fabricate a processor that implements features of the present disclosure. The results of this process can be a mask work, which is then used in a semiconductor manufacturing process to manufacture a processor that implements features of the present disclosure.

[0192] Any of the functions and methods described herein can be implemented in computer programs, software, or firmware incorporated in non-transitory computer-readable storage media for execution by a general purpose computer or a processor. Examples of non-transitory computer-readable storage media include read-only memory (ROM), random-access memory (RAM), registers, cache memory, semiconductor memory devices, magnetic media such as internal hard disks and removable disks, magneto-optical media, and optical media such as CD-ROM disks and digital versatile disks (DVDs).

[0193] It should be understood that many variations are possible based on the disclosure herein. Although features and elements are described above in particular combinations, each feature or element can be used alone without the other features and elements or in various combinations with or without other features and elements.

Claims

1. A system for improving a cardiac ablation procedure, the system comprising: a cloud server comprising a database containing information about previously performed cardiac ablations; a local server communicably coupled to the cloud server via a first network; and a surgical system communicably coupled to the local server via a second network, wherein the cloud server is configured to: receive electrical data and anatomical data of a patient's heart from the surgical system via the local server while performing the cardiac ablation procedure, perform a comparison of the electrical data and anatomical data of the heart to the database information, generate a surgical treatment plan for the heart based on the comparison, wherein the surgical treatment plan comprises a mesh of the heart, transmit the surgical treatment plan to the surgical system via the local server, receive additional electrical data and anatomical data of the heart after performing the ablation procedure, compare the additional electrical data and anatomical data to the treatment plan, based on the comparison, revise the treatment plan for additional ablation.

2. The system of claim 1, wherein one or more parameters of the surgical system are changed in accordance with the treatment plan.

3. The system of claim 1, wherein ablation procedure comprises a plurality of ablations, wherein the cloud server receives the electrical data and anatomical data of the patient's heart from the surgical system after each ablation.

4. The system of claim 1, wherein the cloud server is further configured to: generate an ablation score based on the additional electrical data and anatomical data.

Images