Bootstrapping of Patient-Specific Simulations of Cardiac Electrical Activity
Systems are provided for generating data representing electromagnetic states of a heart for medical, scientific, research, and/or engineering purposes. The systems generate the data based on source configurations such as dimensions of, and scar or fibrosis or pro-arrhythmic substrate location within, a heart and a computational model of the electromagnetic output of the heart. The systems may dynamically generate the source configurations to provide representative source configurations that may be found in a population. For each source configuration of the electromagnetic source, the systems run a simulation of the functioning of the heart to generate modeled electromagnetic output (e.g., an electromagnetic mesh for each simulation step with a voltage at each point of the electromagnetic mesh) for that source configuration. The systems may generate a cardiogram for each source configuration from the modeled electromagnetic output of that source configuration for use in predicting the source location of an arrhythmia.
1 . One or more computing systems for bootstrapping simulations of electromagnetic (EM) output a patient heart of a patient, the one or more computing systems comprising:
one or more computer-readable storage mediums that store computer-executable instructions for controlling the one or more computing systems to:
identify one or more first simulations of EM output a heart, each first simulation based on a heart configuration, each first simulation having an EM output for each of a plurality of simulation steps; and
for the one or more first simulations,
for one or more source locations within a heart,
initialize patient-specific EM output of a patient-specific simulation to the EM output of a simulation step of that first simulation; and
run that patient-specific simulation to generate patient-specific EM output for simulation steps of the patient-specific simulation based on the initialized patient-specific EM output and based on a patient-specific source configuration and that source location, the patient-specific source configuration including parameters derived from the patient heart; and
one or more processors for controlling the one or more computing systems to execute one or more of the computer-executable instructions.
2 . The one or more computing systems of claim 1 wherein the instructions further include instructions for controlling the one or more computing systems to, for each patient-specific simulation, generate a patient-specific cardiogram.
3 . The one or more computing systems of claim 2 wherein the instructions further include instructions for controlling the one or more computing systems to, for each patient-specific simulation, map the patient-specific cardiogram generated based on the patient-specific EM output of that patient-specific simulation to the source location of that patient-specific simulation.
4 . The one or more computing systems of claim 3 wherein the instructions further include instructions for controlling the one or more computing systems to:
access a patient cardiogram of the patient;
identify a patient-specific cardiogram based on similarity to the patient cardiogram; and
output an indication of the source location to which the identified patient-specific cardiogram is mapped as an indication of a patient source location of the patient.
5 . The one or more computing systems of claim 4 wherein the instructions that output display a graphic of a heart along with the indication of the source location.
6 . The one or more computing systems of claim 4 wherein the indication of the source location is output to an ablation therapy device.
7 . The one or more computing systems of claim 3 wherein the instructions further include instructions for controlling the one or more computing systems to perform machine learning training based on training data that includes, for patient-specific simulations, a patient-specific cardiogram of that patient-specific simulation labeled with the source location of that patient-specific simulation.
8 . The one or more computing systems of claim 7 wherein the machine learning training learns parameters of a neural network.
9 . The one or more computing systems of claim 7 wherein the machine learning training learns parameters of a convolutional neural network.
10 . The one or more computing systems of claim 7 wherein the instructions are further for controlling the one or more computing systems to:
access a patient cardiogram of the patient;
identify a patient source location by applying a machine learning algorithm that is trained by the machine learning training to the patient cardiogram; and
output the patient source location.
11 . The one or more computing systems of claim 10 wherein the patient source location is output to an ablation therapy device.
12 . The one or more computing systems of claim 2 wherein a patient-specific cardiogram is generated based on difference between geometry of a first simulation and geometry of the patient heart.
13 . The one or more computing systems of claim 2 wherein a patient-specific cardiogram is generated based on the patient-specific EM output of a patient-specific simulation.
14 . The one or more computing systems of claim 1 wherein at least some of the patient-specific simulations are based on geometry of the patient heart.
15 . The one or more computing systems of claim 1 wherein at least some of the patient-specific simulations are based on electrical characteristics of the patient heart.
16 . The one or more computing systems of claim 1 wherein at least some of the patient-specific simulations are based on geometry of a first simulation.
17 . The one or more computing systems of claim 1 wherein the identification of at least some of the first simulations is based on similarity between the source configurations and the patient-specific source configurations.
18 . A method performed by one or more computing system for bootstrapping generation of a simulated cardiogram, the method comprising:
identifying simulated electromagnetic (EM) output of a step of a simulation of electrical activity of a heart based on a heart configuration, the simulation for generating simulated EM output for a plurality of simulation steps;
initializing patient-specific EM output of a patient-specific simulation to the identified simulated EM output; and
running a patient-specific simulation of electrical activity of the heart of a patient to generate patient-specific EM output for simulation steps of the patient-specific simulation, the patient-specific simulation based on a patient heart configuration of the patient.
19 . The method of claim 18 further comprising repeating the identifying, initializing, and running for each of a plurality simulated EM output of simulations based on different heart configurations.
20 . The method of claim 19 wherein at least some of the patient-specific simulations are based on a source location of an arrhythmia and further comprising, for each of a plurality of simulations, map a patient-specific cardiogram derived from patient-specific EM output of that simulation to the source location of that simulation.
21 . The method of claim 20 further comprising:
accessing a patient cardiogram of the patient;
identifying a patient-specific cardiogram based on similarity to the patient cardiogram; and
outputting an indication of the source location to which the identified patient-specific cardiogram is mapped as an indication of a patient source location of an arrhythmia of the patient.
22 . The method of claim 21 wherein the outputting includes displaying a graphic of a heart along with the indication of the patient source location.
23 . The method of claim 21 wherein the outputting includes outputting an indication of the patient source location to an ablation therapy device.
24 . The method of claim 19 wherein at least some of the patient-specific simulations are based on a source location of an arrhythmia and further comprising training a machine learning algorithm based on training data that includes patient-specific cardiograms derived from the patient-specific EM output of the patient-specific simulations labeled with source locations.
25 . The method of claim of claim 24 wherein the machine learning algorithm includes a convolutional neural network that inputs a cardiogram and outputs a source location.
26 . The method of claim of claim 24 wherein the machine learning algorithm is a neural network that inputs a cardiogram and outputs a source location.
27 . The method of claim 24 further comprising:
accessing a patient cardiogram of the patient;
identifying a patient source location by applying the machine learning algorithm to the patient cardiogram; and
output an indication of the patient source location.
28 . The method of claim 27 wherein the indication of the patient source location is output to an ablation therapy device.
29 . The method of claim 19 wherein at least some of the patient-specific simulations are based on geometry of the heart of a patient.
30 . The method of claim 18 further comprising generating a patient-specific cardiogram based on the patient-specific EM output of the patient-specific simulation.
31 . One or more computer-readable storage mediums that store computer-executable instructions for controlling one or more computing systems to:
identify one or more first simulations of EM output a heart, each first simulation based on a heart configuration, each first simulation having an EM output for each of a plurality of simulation steps; and
for each of a plurality of first simulations and source location of an arrhythmia,
initialize patient-specific EM output of a patient-specific simulation to the EM output of a simulation step of that first simulation; and
run that patient-specific simulation to generate patient-specific EM output for simulation steps of the patient-specific simulation based on the initialized patient-specific EM output and based on a patient-specific source configuration and that source location, the patient-specific source configuration including parameters derived from a patient heart of a patient.
32 . The one or more computer-readable storage mediums of claim 31 the computer-executable instructions are further for controlling the one or more computing systems to control the one or more computing systems to, for each patient-specific simulation, generate a patient-specific cardiogram.
33 . The one or more computer-readable storage mediums of claim 32 wherein the computer-executable instructions are further for controlling the one or more computing systems to:
identify a patient-specific cardiogram based on similarity to a patient cardiogram of the patient; and
output an indication of the source location of the patient-specific simulation that generated the patient-specific EM output from which the identified patient-specific cardiogram was generated.
34 . The one or more computer-readable storage mediums of claim 33 wherein the computer-executable instructions are further for controlling the one or more computing systems to direct treatment of the patient based on the indication of the source location.
35 . The one or more computer-readable storage mediums of claim 34 wherein the treatment is an ablation procedure.
36 . The one or more computer-readable storage mediums of claim 31 wherein at least some of the first simulations are identified based on similarity between the heart configuration of a first simulation and a patient heart configuration of the patient.