Artificial intelligence based cardiac event predictor systems and methods

1. A method for determining cardiac disease risk from electrocardiogram trace data and clinical data, comprising:

receiving electrocardiogram trace data associated with a patient, the electrocardiogram trace data having an electrocardiogram configuration including a plurality of leads and a time interval and comprising, for each lead included in the plurality of leads, voltage data associated with at least a portion of the time interval;

receiving clinical data associated with the patient;

providing the clinical data and at least a portion of the electrocardiogram trace data to a trained machine learning model, the model trained to evaluate the clinical data and the portion of the electrocardiogram trace data with respect to one or more cardiac disease states;

generating, by the trained machine learning model and based on the evaluation, a risk score reflecting a likelihood of the patient being diagnosed with a cardiac disease state within a predetermined period of time from when the electrocardiogram trace data was generated; and

outputting the risk score to at least one of a memory or a display.

2. The method of claim 1 , wherein the clinical data comprises age and gender of the patient.

3. The method of claim 2 , further comprising:

providing the at least a portion of the electrocardiogram trace data to a convolutional neural network; and

providing the age and gender to a boosting model.

4. The method of claim 1 , wherein the trained model further comprises:

training a convolutional neural network on a first plurality of patients and a second plurality of patients, wherein a time between a date of a recorded ECG and a diagnosis of aortic stenosis for each of the first plurality of patients is less than a diagnosis threshold, and a time between a date of a recorded ECG and a diagnosis of aortic stenosis for each of the second plurality of patients is greater than the diagnosis threshold; wherein a time between a date of a recorded ECG and a diagnosis of aortic stenosis that is less than the diagnosis threshold is indicative of an active aortic stenosis at the date of the recorded ECG; and

providing the trained convolutional neural network as the trained model.

5. The method of claim 4 , wherein the trained model further comprises:

refining the trained neural network using only the second plurality of patients, and wherein the diagnosis threshold is selected from a number of days.

6. The method of claim 1 , wherein the trained model is selected based at least in part on a severity of the cardiac disease state the generated risk score represents.

7. The method of claim 6 , wherein the severity includes labels for one or more of normal, mild, moderate, and severe.

8. The method of claim 1 , wherein the predetermined period of time includes times at least one year from when the electrocardiogram trace data was generated.

9. The method of claim 1 , wherein the trained model comprises training data associated with a plurality of clinical sites.

10. The method of claim 9 , wherein the trained model comprises training the model using patient data associated with one site of the plurality of sites and testing the trained model on the remaining sites of the plurality of sites.

11. The method of claim 1 , wherein the outputting the risk score further comprises outputting the risk score to a display of an electrocardiogram monitor, wherein the received electrocardiogram trace data is received from the electrocardiogram monitor.

12. The method of claim 1 , wherein the outputting the risk score further comprises outputting the risk score to an electronic health records management system.

13. The method of claim 1 , wherein the clinical data is selected from demographic data, vitals data, laboratory data, or comorbidities data.

14. The method of claim 13 , wherein the vitals data comprises one or more of body mass index, systolic blood pressure, diastolic blood pressure, heart rate, height, weight, or smoking status.

15. The method of claim 13 , wherein the laboratory data comprises one or more of A1C, bilirubin, blood urea nitrogen, cholesterol, creatine kinase myocardial band, creatinine, C-reactive protein, D-dimer, glucose, high-density lipoprotein, hemoglobin, high-density lipoprotein, lactate dehydrogenase, lymphocytes, potassium, pro B-type natriuretic peptide, sodium, troponin I and T, triglyceride, uric acid, very low density lipoprotein, or estimated glomerular filtration rate.

16. The method of claim 13 , wherein the comorbidities data comprises one or more of heart failure, prior myocardial infarction, diabetes mellitus, chronic obstructive pulmonary disease, renal failure, prior echocardiogram, coronary artery disease, or hypertension.

17. The method of claim 1 , wherein the electrocardiogram trace data comprises ECG data selected from one or more of acute myocardial infarction, atrial fibrillation, atrial flutter, complete block, early repolarization, fascicular block, first-degree atrioventricular block, intraventricular conduction block, left bundle branch block, right bundle branch block, ischemia, left anterior descending artery ischemia, right bundle branch block, low QRS, left ventricular hypertrophy, non-specific ST-T wave, Non-specific T wave, other bradycardia, premature atrial contractions, pacemaker, poor tracing, prior infarction, prior myocardial infarction anterior, prolonged QT, premature ventricular contractions, right axis deviation, second degree atrioventricular block, sinus bradycardia, supraventricular tachycardia, tachycardia, tachyarrhythmia, T inversion, or ventricular tachycardia.

18. The method of claim 1 , further comprising gathering the clinical data and electrocardiogram trace data based at least in part on a presence of ICD codes associated with the patient.

19. The method of claim 1 , wherein the electrocardiogram trace data comprises at least 8 leads.

20. The method of claim 1 , wherein the electrocardiogram trace data is sampled at 250 hz or 500 hz.

21. The method of claim 1 , wherein the trained model is a composite model, and wherein training the trained composite model comprises:

generating a patient timeline for each patient;

anchoring each respective patient timeline to a date of occurrence of an echocardiogram; and

labeling each respective patient as having a positive or negative ECG based at least in part on a date of an ECG with respect to the date of occurrence of the echocardiogram.

22. The method of claim 21 , further comprises excluding patients from training after a censoring event is detected in the patient timeline.

23. The method of claim 1 , wherein the cardiac disease state is one of aortic stenosis, aortic regurgitation, mitral stenosis, mitral regurgitation, tricuspid regurgitation, reduced ejection fraction, increased interventricular septal thickness, cardiac amyloidosis, or stroke.

24. A system comprising:

a computer including a processing device, the processing device configured to:

receive electrocardiogram trace data associated with a patient, the electrocardiogram trace data having an electrocardiogram configuration including a plurality of leads and a time interval and comprising, for each lead included in the plurality of leads, voltage data associated with at least a portion of the time interval;

receive clinical data associated with the patient;

provide the clinical data and at least a portion of the electrocardiogram trace data to a trained machine learning model, the model trained to evaluate the clinical data and the portion of the electrocardiogram trace data with respect to one or more cardiac disease states;

generate, by the trained machine learning model and based on the evaluation, a risk score reflecting a likelihood of the patient being diagnosed with a cardiac disease state within a predetermined period of time from when the electrocardiogram trace data was generated; and

output the risk score to at least one of a memory or a display.

25. The system of claim 24 , wherein the clinical data comprises age and gender of the patient.

26. The system of claim 24 , wherein the cardiac disease state is one of aortic stenosis, aortic regurgitation, mitral stenosis, mitral regurgitation, tricuspid regurgitation, reduced ejection fraction, increased interventricular septal thickness, cardiac amyloidosis, or stroke.

27. A non-transitory computer readable medium, comprising instructions for causing a computer to:

receive electrocardiogram trace data associated with a patient, the electrocardiogram trace data having an electrocardiogram configuration including a plurality of leads and a time interval and comprising, for each lead included in the plurality of leads, voltage data associated with at least a portion of the time interval;

receive clinical data associated with the patient;

provide the clinical data and at least a portion of the electrocardiogram trace data to a trained machine learning model, the model trained to evaluate the clinical data and the portion of the electrocardiogram trace data with respect to one or more cardiac disease states;

generate, by the trained machine learning model and based on the evaluation, a risk score reflecting a likelihood of the patient being diagnosed with a cardiac disease state within a predetermined period of time from when the electrocardiogram trace data was generated; and

output the risk score to at least one of a memory or a display.

28. The non-transitory computer readable medium of claim 27 , wherein the clinical data comprises age and gender of the patient.

29. The non-transitory computer readable medium of claim 27 , wherein the cardiac disease state is one of aortic stenosis, aortic regurgitation, mitral stenosis, mitral regurgitation, tricuspid regurgitation, reduced ejection fraction, increased interventricular septal thickness, cardiac amyloidosis, or stroke.