Electrocardiogram processing system for detecting and/or predicting cardiac events
Systems and methods are provided for analyzing electrocardiogram (ECG) data of a patient using a substantial amount of ECG data. The systems receive ECG data from a sensing device positioned on a patient such as one or more ECG leads/electrodes that may be integrated in a smart device. The system may include an application that communicates with an ECG platform running on a server(s) that processes and analyzes the ECG data, e.g., using neural networks, to detect and/or predict various abnormalities, conditions and/or descriptors. The system may also determine a confidence score corresponding to the abnormalities, conditions and/or descriptors. The processed ECG data is used to generate a graphic user interface that is communicated from the server(s) to a computer for display in a user-friendly and interactive manner with enhanced accuracy.
1. A computerized-system for analyzing electrocardiogram (ECG) data of a patient, the computerized-system configured to:
analyze the ECG data to determine a presence of atrial fibrillation in the ECG data, the ECG data sampled over a plurality of time points;
if the atrial fibrillation is determined to be present based on the analysis of the ECG data, generate information to identify the presence of the atrial fibrillation for display; and
if the atrial fibrillation is determined not to be present based on the analysis of the ECG data, further analyze the ECG data, wherein the analysis comprises:
perform delineation on the ECG data using a delineation neural network to generate delineation output values indicative of a likelihood of the presence of one or more waves at each time point of the plurality of time points;
determine a plurality of beats based on the values; and
perform classification on the ECG data and the plurality of beats to generate classification output values indicative of a likelihood of a presence of one or more normal beats, premature atrial complexes (PAC), or premature ventricular complexes (PVC),
wherein the further analysis comprises:
apply the ECG data, delineation output values, and classification output values to a risk analyzer, the risk analyzer comprising one or more neural networks trained to determine a risk of atrial fibrillation; and
determine a risk score using the risk analyzer, the risk score based on at least the ECG data, the delineation output values, and the classification output values, the risk score indicative of future atrial fibrillation for the patient, the future atrial fibrillation being at a time in the future after the plurality of time points for which the ECG data was sampled.
2. The computerized-system of claim 1 , the computerized-system further configured to:
extract a plurality of first features from the plurality of beats; and
determine a first risk score based on the plurality of first features, the first risk score indicative of future atrial fibrillation for the patient.
3. The computerized-system of claim 2 , wherein performing delineation on the ECG data determines a plurality of QRS onset values.
4. The computerized-system of claim 2 , the computerized-system further configured to:
determine a plurality of timestamps corresponding to the plurality of beats;
determine a plurality of heart rate values based on the plurality of timestamps;
determine a matrix based on the plurality of timestamps and the plurality of heart rate values;
generate a graphical representation of the matrix; and
process the graphical representation of the matrix using at least one neural network to determine a second risk score indicative of future atrial fibrillation for the patient.
5. The computerized-system of claim 4 , wherein the at least one neural network is at least one deep neural network (DNN).
6. The computerized-system of claim 4 , the computerized-system further configured to:
determine patient information corresponding to the patient, the patient information comprising one or more of age and sex of the patient;
determine a plurality of second features of the ECG data indicative of atrial fibrillation; and
apply the plurality of second features to a classifier to determine a third risk score indicative of future atrial fibrillation for the patient.
7. The computerized-system of claim 6 , wherein the classifier is one or more of DNN, logistic regression and a Random Forest.
8. The computerized-system of claim 6 , wherein the risk score is based on one or more of the first risk score, the second risk score and the third risk score.
9. The computerized-system of claim 6 , the computerized-system further configured to:
determine an average of the first risk score, the second risk score, and the third risk score to determine the risk score.
10. The computerized-system of claim 6 , the computerized-system further configured to:
apply the first risk score, the second risk score, and the third risk score to a first neural network trained to determine the risk score.
11. The computerized-system of claim 1 , wherein the risk score is indicative of a likelihood of atrial fibrillation in a set period of time, and the computerized-system is further configured to:
cause an ECG recorder to capture second ECG data during the set period of time.
12. A method for analyzing electrocardiogram (ECG) data of a patient, the method comprising:
analyzing the ECG data to determine a presence of atrial fibrillation in the ECG data, the ECG data sampled over a plurality of time points;
if the atrial fibrillation is determined to be present based on the analysis of the ECG data, generating information to identify the presence of the atrial fibrillation for display; and
if the atrial fibrillation is determined not to be present based on the analysis of the ECG data, further analyzing the ECG data, wherein the analysis comprises:
performing delineation on the ECG data using a delineation neural network to generate delineation output values indicative of a likelihood of the presence of one or more waves at each time point of the plurality of time points;
determining a plurality of beats based on the values; and
performing classification on the ECG data and the plurality of beats to generate classification output values indicative of a likelihood of a presence of one or more normal beats, premature atrial complexes (PAC), or premature ventricular complexes (PVC),
wherein the further analysis comprises:
apply the ECG data, delineation output values, and classification output values to a risk analyzer, the risk analyzer comprising one or more neural networks trained to determine a risk of atrial fibrillation; and
determining a risk score using the risk analyzer, the risk score based on at least the ECG data, the delineation output values, and the classification output values, the risk score indicative of future atrial fibrillation for the patient, the future atrial fibrillation being at a time in the future after the plurality of time points for which the ECG data was sampled.
13. The method of claim 12 , further comprising:
extracting a plurality of first features from the plurality of beats; and
determining a first risk score based on the plurality of first features, the first risk score indicative of future atrial fibrillation for the patient.
14. The method of claim 13 , wherein performing delineation on the ECG data determines a plurality of QRS onset values.
15. The method of claim 13 , further comprising:
determining a plurality of timestamps corresponding to the plurality of beats;
determining a plurality of heart rate values based on the plurality of timestamps;
determining a matrix based on the plurality of timestamps and the plurality of heart rate values;
generating a graphical representation of the matrix; and
processing the graphical representation of the matrix using at least one neural network to determine a second risk score indicative of future atrial fibrillation for the patient.
16. The method of claim 15 , wherein the at least one neural network is at least one deep neural network (DNN).
17. The method of claim 15 , further comprising:
determining patient information corresponding to the patient, the patient information comprising one or more of age and sex of the patient;
determining a plurality of second features of the ECG data indicative of atrial fibrillation; and
applying the plurality of second features to a classifier to determine a third risk score indicative of future atrial fibrillation for the patient.
18. The method of claim 17 , wherein the classifier is one or more of DNN, logistic regression and a Random Forest.
19. The method of claim 17 , wherein the risk score is based on one or more of the first risk score, the second risk score and the third risk score.
20. The method of claim 17 , further comprising:
determining an average of the first risk score, the second risk score, and the third risk score to determine the risk score.
21. The method of claim 17 , further comprising:
applying the first risk score, the second risk score, and the third risk score to a first neural network trained to determine the risk score.
22. The method of claim 12 , wherein the risk score is indicative of a likelihood of atrial fibrillation in a set period of time, and further comprising:
causing an ECG recorder to capture second ECG data during the set period of time.
23. A non-transitory computer-readable memory medium configured to store instructions thereon that, when loaded by at least one processor, cause the at least one processor to:
analyze ECG data of a patient to determine a presence of atrial fibrillation in the ECG data, the ECG data sampled over a plurality of time points;
if the atrial fibrillation is determined to be present based on the analysis of the ECG data, generate information to identify the presence of the atrial fibrillation for display; and
if the atrial fibrillation is determined not to be present based on the analysis of the ECG data, further analyze the ECG data, wherein the analysis comprises:
perform delineation on the ECG data using a delineation neural network to generate delineation output values indicative of a likelihood of the presence of one or more waves at each time point of the plurality of time points;
determine a plurality of beats based on the values; and
perform classification on the ECG data and the plurality of beats to generate classification output values indicative of a likelihood of a presence of one or more normal beats, premature atrial complexes (PAC), or premature ventricular complexes (PVC),
wherein the further analysis comprises:
apply the ECG data, delineation output values, and classification output values to a risk analyzer, the risk analyzer comprising one or more neural networks trained to determine a risk of atrial fibrillation; and
determine a risk score using the risk analyzer, the risk score based on at least the ECG data, the delineation output values, and the classification output values, the risk score indicative of future atrial fibrillation for the patient, the future atrial fibrillation being at a time in the future after the plurality of time points for which the ECG data was sampled.
24. The non-transitory computer-readable memory medium of claim 23 , further configured to cause the at least one processor to:
extract a plurality of first features from the plurality of beats; and
determine a first risk score based on the plurality of first features, the first risk score indicative of future atrial fibrillation for the patient.
25. The non-transitory computer-readable memory medium of claim 24 , further configured to cause the at least one processor to:
determine a plurality of timestamps corresponding to the plurality of beats;
determine a plurality of heart rate values based on the plurality of timestamps;
determine a matrix based on the plurality of timestamps and the plurality of heart rate values;
generate a graphical representation of the matrix; and
process the graphical representation of the matrix using at least one neural network to determine a second risk score indicative of future atrial fibrillation for the patient.
26. The non-transitory computer-readable memory medium of claim 25 , further configured to cause the at least one processor to:
determine patient information corresponding to the patient, the patient information comprising one or more of age and sex of the patient;
determine a plurality of second features of the ECG data indicative of atrial fibrillation; and
apply the plurality of second features to a classifier to determine a third risk score indicative of future atrial fibrillation for the patient.
27. The non-transitory computer-readable memory medium of claim 26 , wherein the risk score is based on one or more of the first risk score, the second risk score and the third risk score.