METHOD AND APPARATUS FOR CARDIAC ARRHYTHMIA CLASSIFICATION USING TEMPLATE BAND-BASED MORPHOLOGY ANALYSIS

• US 20090005826A1
• Filed: 09/08/2008
• Published: 01/01/2009
• Est. Priority Date: 01/20/2005
• Status: Active Grant

• Current Assignee: Cardiac Pacemakers Incorporated (Boston Scientific Corporation)
• Original Assignee: Cardiac Pacemakers Incorporated (Boston Scientific Corporation)
• Inventors: Li, Dan
• Granted Patent: US 8,244,348 B2
• Time in Patent Office: Days
• US Class Current: 607/5
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/363 Detecting tachycardia or bradycardia
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

An implantable cardioverter/defibrillator (ICD) includes a tachyarrhythmia detection and classification system that classifies tachyarrhythmias based on a morphological analysis of arrhythmic waveforms and a template waveform. Correlation coefficients each computed between morphological features of an arrhythmic waveform and morphological features of the template waveform provide for the basis for classifying the tachyarrhythmia. In one embodiment, a correlation analysis takes into account the uncertainty associated with the production of the template waveform by using a template band that includes confidence intervals.

Claims

1. A system for classifying tachyarrhythmias, the system comprising:

a feature vector generation circuit configured to produce a template feature vector (a) using a plurality of template morphological features on a plurality of template waveforms each associated a template heart beat of a known type cardiac rhythm, an arrhythmic feature vector (b) using a plurality of arrhythmic morphological features of an arrhythmic waveform associated with an arrhythmic heart beat sensed during a tachyarrhythmia, and a maximum feature vector (a_max) and a minimum feature vector (a_min) using the plurality of template morphological features, the plurality of arrhythmic morphological features, the maximum feature vector (a_max) and the minimum feature vector (a_min) being within a template band related to predetermined confidence levels of the plurality of template morphological features;

a correlation computing circuit coupled to the feature vector generation circuit, the correlation computing circuit configured to calculate a mean feature correlation coefficient (Fcc_mean) using the template feature vector (a) and the arrhythmic feature vector (b), a maximum feature correlation coefficient (Fcc_max) using the maximum feature vector (a_max) and the arrhythmic feature vector (b), and a minimum feature correlation coefficient (Fcc_min) using the minimum feature vector (a_min) and the arrhythmic feature vector (b); and

a beat classification circuit coupled to the correlation analysis circuit, the beat classification circuit configured to classify the arrthythmic heart beat using the mean feature correlation coefficient (Fcc_mean), the maximum feature correlation coefficient (Fcc_max), the minimum feature correlation coefficient (Fcc_min), and at least one predetermined correlation threshold.

2. The system of claim 1, wherein the beat classification circuit comprises a decision circuit configured to calculate a score for the arrhythmic heart beat using the mean feature correlation coefficient (FCC_mean), the maximum feature correlation coefficient (Fcc_max), the minimum feature correlation coefficient (Fcc_min), and the predetermined correlation threshold, the score indicative of a probability of the tachyarrhythmia being a first type tachyarrhythmia.

3. The system of claim 2, comprising an episode classification circuit coupled to the beat classification circuit, the episode classification circuit configured to classify the tachyarrhythmia using the scores produced for a plurality of arrhythmic heart beats sensed during the tachyarrhythmia.

4. The system of claim 3, wherein the episode classification circuit comprises:

a score averaging circuit configured to calculate an episode score being an average of the scores produced for the plurality of arrhythmic heart beats; and

a comparator configured to compare the episode score to a predetermined classification threshold and classify the tachyarrhythmia as the first type tachyarrhythmia in response to the episode score exceeding the predetermined classification threshold.

5. The system of claim 1, wherein feature vector generation circuit comprises:

a template band construction circuit to produce the template feature vector (a) and a template standard deviation vector (σ

) using the plurality of template morphological features;

an arrhythmic feature vector generation circuit configured to produce the arrhythmic feature vector (b) using the plurality of arrhythmic morphological features;

a deviation vector generation circuit configured to produce a maximum deviation vector (x_max) and a minimum deviation vector (x_min) based on at least the template feature (A) vector, the arrhythmic feature vector (b), and the template standard deviation vector (a), the maximum deviation vector (x_max) and the minimum deviation vector (x_min) indicative of morphological variations among the plurality of template morphological features; and

a boundary feature vector generation circuit configured to produce the maximum feature vector (a_max) by adding the maximum deviation vector (x_max) to the template feature vector (a) and produce the minimum feature vector (a_min) by adding the minimum deviation vector (x_min) to the template feature vector (a).

6. The system of claim 5, wherein the template band construction circuit comprises:

a template vector generator configured to produce a plurality of template feature vectors each associated with one template heart beat of the plurality of template heart beats;

a mean vector calculator configured to produce the template feature vector (a) being a mean vector of the plurality of template feature vectors; and

a standard deviation vector calculator configured to produce the template standard deviation vector (σ

) being a standard deviation vector of the plurality of template feature vectors.

7. The system of claim 1, comprising an implantable cardioveitor/defibrillator (ICD) including the feature vector generation circuit, the correlation computing circuit, and the beat classification circuit.

8. The system of claim 7, wherein the ICD comprises:

a sensing circuit configured to sense an atrial electrogram and a ventricular electrogram;

a rate detection circuit configured to detect to an atrial rate using the atrial electrogram and a ventricular rate using the ventricular electrogram; and

a tachyarrhythmia detection circuit configured to detect the tachyarrhythmia using at least one of the atrial rate and the ventricular rate.

9. The system of claim 8, wherein the ICD comprises a rhythm classification circuit configured to classify the detected tachyarrhythmia as a 1:

1 tachyarrhythmia in response to a determination that the atrial rate and the ventricular rate are substantially equal during the detected tachyarrhythmia, and wherein the feature vector generation circuit is adapted to produce the arrhythmic feature vector (b) based on a plurality of arrhythmic morphological features of an arrhythmic waveform associated with an arrhythmic heart beat sensed during the detected tachyarrhythmia classified as the 1;

1 tachyarrhythmia.

10. A method for classifying tachyarrhythmias, the method comprising:

producing a template feature vector (a) and a template standard deviation vector (σ

) based on template morphological features on a plurality of template waveforms each associated with a heart beat of a known type cardiac rhythm;

producing an arrhythmic feature vector (b) based on arrhythmia morphological features on an arrhythmic waveform associated with an arrhythmic heart beat sensed during a tachyarrhythmia;

producing a maximum deviation vector (x_max) and a minimum deviation vector (x_min) based on at least the template feature vector (a), the arrhythmic feature vector (b), and the template standard deviation vector (σ

), the maximum deviation vector (x_max) and the minimum deviation vector (a_min) indicative of morphological variations among the plurality of template morphological features;

producing a maximum feature vector (a_max) by adding the maximum deviation vector (x_max) to the template feature vector (g) and a minimum feature vector (a_min) by adding the minimum deviation vector (x_min) to the template feature vector (a); and

computing a mean feature correlation coefficient (Fcc_mean) based on the template feature vector (a) and the arrhythmic feature vector (b), a maximum feature correlation coefficient (Fcc_max) based on the maximum feature vector (a_max) and the arrhythmic feature vector (b), and a minimum feature correlation coefficient (Fcc_min) based on the minimum feature vector (a_min) and the arrhythmic feature vector (b).

11. The method of claim 10, further comprising classifying the arrhythmic heart beat based on the mean feature correlation coefficient (Fcc_mean), the maximum feature correlation coefficient (Fcc_max), the minimum feature correlation coefficient (Fcc_min), and at least one predetermined correlation threshold.

12. The method of claim 11, wherein classifying the arrhythmic heart beat comprises computing a score for the arrhythmic heart beat based on the mean feature correlation coefficient (Fcc_mean), the maximum feature correlation coefficient (Fcc_max), the minimum feature correlation coefficient (Fcc_min), and a predetermined correlation threshold, the score indicative a probability of the tachyarrhythmia being the first type tachyarrhythmia.

13. The method of claim 12, wherein classifying the arrhythmic heart beat further comprises:

setting the score to 1 if the maximum feature correlation coefficient (Fcc_max) is below the predetermined correlation threshold; and

setting the score to 0 if the minimum feature correlation coefficient (Fcc_min) exceeds the predetermined correlation threshold.

14. The method of claim 12, further comprising classifying the tachyarrhythmia based on the scores produced for a plurality of arrhythmic heart beats of the tachyarrhythmia.

15. The method of claim 14, wherein classifying the tachyarrhythmia comprises:

calculating an episode score being an average of the scores produced for the plurality of arrhythmic heart beats;

comparing the episode score to a predetermined classification threshold; and

classifying the tachyarrhythmia as the first type tachyarrhythmia if the episode score exceeds the predetermined classification threshold.

16. The method of claim 15, wherein the known rhythm is a normal sinus rhythm, and the first type tachyarrhythmia is a supraventricular tachyarrhythmia.

17. The method of claim 10, wherein producing the template feature vector (a) and a template standard deviation vector (σ

) comprises;

receiving a plurality of feature vectors each associated with one template waveform of the plurality of template waveforms;

producing the template feature vector (a) being a mean vector of the plurality of feature vectors; and

producing the template standard deviation vector (σ

) being a standard deviation vector of the plurality of feature vectors.

18. The method of claim 17, wherein producing the arrhythmic feature vector (b) comprises:

receiving the arrhythmic waveform;

extracting the morphological features from the arrhythmic waveform; and

producing the arrhythmic feature vector (b) based on the plurality of morphological features.

19. The method of claim 10, comprising:

sensing an atrial electrogram and a ventricular electrogram;

detecting an atrial rate using the atrial electrogram;

detecting a ventricular rate using the ventricular electrogram; and

detecting the tachyarrhythmia using at least one of the atrial rate and the ventricular rate.

20. The method of claim 19, wherein the tachyarrhythmia is a 1:

1 tachyarrhythmia, and comprising classifying the detected tachyarrhythmia as a 1;

1 tachyarrhythmia when the atrial rate and the ventricular rate are substantially equal.