METHOD AND APPARATUS FOR ALGORITHM FUSION OF HIGH-RESOLUTION ELECTROCARDIOGRAMS

• US 20050004481A1
• Filed: 07/01/2003
• Published: 01/06/2005
• Est. Priority Date: 07/01/2003
• Status: Active Grant

• Current Assignee: GE Medical Systems Information Technologies Incorporated (GE Healthcare Technologies, Inc.)
• Original Assignee: GE Medical Systems Information Technologies Incorporated (GE Healthcare Technologies, Inc.)
• Inventors: BRODNICK, DONALD EUGENE, XUE, JOEL Q., ELKO, PAUL P.
• Granted Patent: US 7,142,907 B2
• Time in Patent Office: Days
• US Class Current: 600/509
• CPC Class Codes: A61B 5/02405 Determining heart rate variability
• CPC Class Codes: A61B 5/349 Detecting specific parameters of the electrocardiograph cycle
• CPC Class Codes: A61B 5/355 Detecting T-waves
• CPC Class Codes: A61B 5/36 Detecting PQ interval, PR interval or QT interval

Abstract

A system and a method for noninvasive ECG detection and diagnosis. The system comprises a multiplicity of electrodes applied to a patient, a data acquisition system for acquiring high-resolution ECG data from the patient; and a processor programmed to process the acquired data in accordance with two or more different ECG analysis algorithms, and then derive a prediction score for a particular clinical end point as a function of the respective results of those ECG analysis algorithms.

Claims

1. A method for noninvasive ECG detection and diagnosis, comprising the following steps:

acquiring high-resolution ECG data from a patient;

processing said acquired data in accordance with two or more different ECG analysis algorithms; and

deriving a prediction score for a particular clinical end point as a function of the respective results of said two or more ECG analysis algorithms.

2. The method as recited in claim 1, further comprising the step of training a predictive model with clinically confirmed data for both input and output, wherein said predictive model is used for said prediction score derivation.

3. The method as recited in claim 1, wherein said particular clinical end point is sudden cardiac death.

4. The method as recited in claim 1, wherein said particular clinical end point is sustained ventricular tachycardia.

5. The method as recited in claim 1, wherein said particular clinical end point is ischemia.

6. The method as recited in claim 1, wherein said ECG analysis algorithms comprise late potential, T wave alternans and QT dynamicity algorithms.

7. The method as recited in claim 1, wherein said ECG analysis algorithms comprise QT dispersion and intra-QRS algorithms.

8. The method as recited in claim 1, wherein said ECG analysis algorithms comprise heart rate variability, QT dispersion and QT dynamicity algorithms.

9. The method as recited in claim 1, wherein said data acquisition step uses the X, Y and Z leads of a Frank lead system.

10. The method as recited in claim 9, wherein one of said ECG analysis algorithms is a T wave alternans algorithm comprising the following steps:

dividing the high-resolution ECG data into even beat and odd beat groups;

averaging beats in said even and odd beat groups separately;

determining the variance of T wave morphology of all even beats for each of said X, Y, Z leads;

determining the variance of T wave morphology of all odd beats for each of said X, Y, Z leads; and

determining the variance of T wave morphology between even and odd averaged beats for each of said X, Y, Z leads.

11. The method as recited in claim 9, wherein said data acquisition step also uses leads of a standard 12-lead system.

12. The method as recited in claim 11, wherein one of said ECG analysis algorithms is a QT dispersion algorithm comprising the following steps:

computing a respective QT dispersion value for each of a multiplicity of successive segments of 12-lead ECGs; and

computing an mean or median value of said QT dispersion values.

13. A system for noninvasive ECG detection and diagnosis, comprising:

a multiplicity of electrodes applied to a patient, a data acquisition system for acquiring high-resolution ECG data from the patient; and

a processor programmed to process said acquired data in accordance with two or more different ECG analysis algorithms, and derive a prediction score for a particular clinical end point as a function of the respective results of said two or more ECG analysis algorithms.

14. The system as recited in claim 13, wherein said processor is programmed to use a predictive model for said prediction score derivation, said predictive model being trained with clinically confirmed data for both input and output.

15. The system as recited in claim 13, wherein said particular clinical end point is sudden cardiac death.

16. The system as recited in claim 13, wherein said particular clinical end point is sustained ventricular tachycardia.

17. The system as recited in claim 13, wherein said particular clinical end point is ischemia.

18. The system as recited in claim 13, wherein said ECG analysis algorithms comprise late potential, T wave alternans and QT dynamicity algorithms.

19. The system as recited in claim 13, wherein said ECG analysis algorithms comprise QT dispersion and intra-QRS algorithms.

20. The system as recited in claim 13, wherein said ECG analysis algorithms comprise heart rate variability, QT dispersion and QT dynamicity algorithms.

21. The system as recited in claim 13, wherein said data acquisition system comprises the X, Y and Z leads of a Frank lead system.

22. The system as recited in claim 21, wherein one of said ECG analysis algorithms executed by said processor is a T wave alternans algorithm comprising the following steps:

dividing the high-resolution ECG data into even beat and odd beat groups;

averaging beats in said even and odd beat groups separately;

determining the variance of T wave morphology of all even beats for each of said X, Y, Z leads;

determining the variance of T wave morphology of all odd beats for each of said X, Y, Z leads; and

determining the variance of T wave morphology between even and odd averaged beats for each of said X, Y, Z leads.

23. The system as recited in claim 21, wherein said data acquisition system further comprises leads of a standard 12-lead system applied to the patient.

24. The system as recited in claim 23, wherein one of said ECG analysis algorithms executed by said processor is a QT dispersion algorithm comprising the following steps:

computing a respective QT dispersion value for each of a multiplicity of successive segments of 12-lead ECGs; and

computing an mean or median value of said QT dispersion values.

25. A method for noninvasive ECG detection and diagnosis, comprising the following steps:

(a) acquiring high-resolution ECG data from a patient using the X, Y and Z leads of a Frank lead system;

(b) dividing the high-resolution ECG data into even beat and odd beat groups;

(c) averaging beats in said even and odd beat groups separately;

(d) determining the variance of T wave morphology of all even beats for each of said X, Y, Z leads;

(e) determining the variance of T wave morphology of all odd beats for each of said X, Y, Z leads;

(f) determining the variance of T wave morphology between even and odd averaged beats for each of said X, Y, Z leads; and

(g) determining T wave alternans as a function of the results of steps (d)-(f) by analysis of variance.

26. A system for noninvasive ECG detection and diagnosis, comprising:

a multiplicity of electrodes applied to a patient, a data acquisition system for acquiring high-resolution ECG data from the patient using the X, Y and Z leads of a Frank lead system; and

a processor programmed to process said acquired data in accordance with a T wave alternans algorithm comprising the following steps;

(a) dividing the high-resolution ECG data into even beat and odd beat groups;

(b) averaging beats in said even and odd beat groups separately;

(c) determining the variance of T wave morphology of all even beats for each of said X, Y, Z leads;

(d) determining the variance of T wave morphology of all odd beats for each of said X, Y, Z leads;

(e) determining the variance of T wave morphology between even and odd averaged beats for each of said X, Y, Z leads; and

(f) determining T wave alternans as a function of the results of steps (c)-(e) by analysis of variance.