Method and apparatus for rate-dependent morphology-based cardiac arrhythmia classification

• US 20060281998A1
• Filed: 06/13/2005
• Published: 12/14/2006
• Est. Priority Date: 06/13/2005
• 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 9,314,210 B2
• Time in Patent Office: Days
• US Class Current: 600/516
• 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 medical device includes a tachyarrhythmia detection and classification system that classifies tachyarrhythmias based on a morphological analysis of template and arrhythmic waveforms. The morphological analysis takes effect of heart rate on the morphological characteristics of the template and arrhythmic waveforms into consideration. Correlation between morphological features of the template waveform and corresponding morphological features of an arrhythmic waveform provides for the basis for classifying the tachyarrhythmia. In one embodiment, corresponding morphological features are extracted from the template and arrhythmic waveforms at locations determined by the heart rate associated with a detected arrhythmia episode. In another embodiment, weighting factors each being a function of the heart rate are applied to the template and arrhythmic morphological features before a correlation coefficient is calculated.

Claims

1. A system for classifying cardiac arrhythmias, the system comprising:

a template waveform input to receive template data representative of a template waveform associated with a template heart beat of a known cardiac rhythm;

an arrhythmic waveform input to receive arrhythmic data representative of an arrhythmic waveform associated with an arrhythmic heart beat of an arrhythmia episode;

an arrhythmic heart rate input to receive an arrhythmic heart rate parameter representative of an arrhythmic heart rate associated with the arrhythmic waveform;

a rate-dependent feature locator coupled to the template waveform input, the arrhythmic waveform input, and the arrhythmic heart rate parameter input, the rate-dependent feature locator adapted to determine morphological feature locations using at least the arrhythmic heart rate parameter, the morphological feature locations including template feature locations of a plurality of template morphological features on the template waveform and arrhythmic feature locations of a plurality of corresponding arrhythmic morphological features on the arrhythmic waveform; and

a feature extracting module coupled to the rate-dependent feature locator, the feature extracting module adapted to extract the plurality of template morphological features from the template waveform and the plurality of corresponding arrhythmic morphological features from the arrhythmic waveform based on the morphological feature locations.

2. The system of claim 1, further comprising:

a correlation analyzer, coupled to the feature extracting module, to produce at least one correlation coefficient representative of a correlation between the plurality of template morphological features and the plurality of corresponding arrhythmic morphological features; and

an arrhythmia classification module, coupled to the correlation analyzer, to classify the arrhythmia episode based on the at least one correlation coefficient.

3. The system of claim 1, further comprising an arrhythmic heart rate estimator coupled to the arrhythmic heart rate input, the arrhythmic heart rate estimator adapted to produce the arrhythmic heart rate parameter by estimating an instantaneous heart rate associated with the arrhythmic heart beat.

4. The system of claim 3, wherein the arrhythmic heart rate estimator is further adapted to produce the arrhythmic heart rate parameter by mapping the estimated instantaneous heart rate to one heart rate range of a plurality of predetermined heart rate ranges.

5. The system of claim 1, further comprising a template heart rate input to receive the template heart rate parameter representative of a template heart rate associated with the template waveform, and wherein the rate-dependent feature locator is further coupled to the template heart rate parameter input and adapted to determine the morphological feature locations using at least the arrhythmic heart rate parameter and the template heart rate parameter.

6. The system of claim 5, further comprising:

a template waveform storage device, coupled to the template waveform input, to store the template data including a digitized template waveform having approximately 64 to 320 sample points; and

a template generator, coupled to the template waveform storage device, to produce the digitized template waveform by sampling a cardiac signal at a sampling frequency in a range of approximately 200 to 1,000 Hz.

7. The system of claim 5, further comprising:

a template waveform storage device, coupled to the template waveform input, to store the template data;

a template generator, coupled to the template waveform storage device, to produce the template data;

an anchor point locator coupled to the template generator, the anchor point locator adapted to locate a plurality of anchor points on the template waveform; and

a waveform divider coupled to the template generator, the anchor point locator, and the template waveform storage device, the waveform divider adapted to divide the template waveform into a plurality of template waveform segments using one or more anchor points of the plurality of anchor points.

8. The system of claim 7, wherein the rate-dependent feature locator comprises a template feature locator adapted to determine the template feature locations using the arrhythmic heart rate parameter, the template heart rate parameter, and one or more selected anchor points of the plurality of anchor points.

9. The system of claim 7, wherein the anchor point locator comprises a turning point locator to locate a first turning point associated with a start of a QRS complex and a second turning point associated with an end of the QRS complex, and the waveform divider is adapted to divide the template waveform into a pre-QRS segment, a QRS segment, and a post-QRS segment using the first and second turning points.

10. The system of claim 9, wherein the rate-dependent feature locator comprises a template feature locator adapted to select the first and second turning points as two of the template feature locations.

11. The system of claim 10, wherein the template feature locator is adapted to select one or more pre-QRS template feature locations each having a first timing relationship with the first turning point from the pre-QRS segment and one or more post-QRS template feature locations each having a second timing relationship with the second turning point from the post-QRS segment, the first timing relationship being a first function of the arrhythmic heart rate parameter, the second timing relationship being a second function of the arrhythmic heart rate parameter.

12. The system of claim 10, wherein the template feature locator is adapted to select a plurality of QRS template feature locations from the QRS segment, the selected QRS template feature locations being a function of the arrhythmic heart rate parameter.

13. The system of claim 8, wherein the rate-dependent feature locator comprises:

a fiducial point locator adapted to locate a template fiducial point on the template waveform and a corresponding arrhythmic fiducial point on the arrhythmic waveform;

a feature timing module to measure feature time intervals each between one template morphological feature of the plurality of template morphological features and the template fiducial point; and

an arrhythmic feature locator adapted to determine the arrhythmic feature locations on the arrhythmic waveform using the measured time intervals each as a time interval between one arrhythmic morphological feature of the plurality of corresponding arrhythmic morphological features and the arrhythmic fiducial point.

14. A method for classifying cardiac arrhythmias, the method comprising:

detecting an arrhythmia episode;

producing a heart rate parameter representative of an arrhythmic heart rate associated with the arrhythmic episode;

determining morphological feature locations using at least the heart rate parameter, the morphological feature locations including template feature locations of a plurality of template morphological features on a template waveform associated with a template heart-beat of a known cardiac rhythm and arrhythmic feature locations of a plurality of corresponding arrhythmic morphological features on an arrhythmic waveform associated with an arrhythmic heart beat of the arrhythmia episode; and

extracting the plurality of template morphological features from the template waveform and the plurality of corresponding arrhythmic morphological features from the arrhythmic waveform based on the morphological feature locations.

15. The method of claim 14, wherein producing the heart rate parameter comprises:

estimating an instantaneous heart rate associated with the arrhythmic waveform; and

mapping the estimated instantaneous heart rate to a heart rate range of a plurality of predetermined heart rate ranges.

16. The method of claim 14, further comprising:

producing template data representative of the template waveform; and

storing the template data,

wherein producing the template data comprises locating a plurality of anchor points on the template waveform and dividing the template waveform into a plurality of template waveform segments using one or more anchor points of the plurality of anchor points.

17. The method of claim 16, wherein determining the morphological feature locations comprises:

selecting one or more locations of the anchor points as one or more locations of the template feature locations; and

selecting a plurality of additional locations of the template feature locations based on timing relationships each between one of the additional locations and one of the one or more locations of the anchor points, the timing relationships each being a function of the arrhythmic heart rate parameter.

18. The method of claim 16, wherein locating the plurality of anchor points comprises locating a first turning point associated with a start of a QRS complex and a second turning point associated with an end of the QRS complex, and dividing the template waveform into a plurality of template waveform segments comprises dividing the template waveform into a pre-QRS segment, a QRS segment, and a post-QRS segment using the first and second turning points.

19. The method of claim 18, wherein determining the morphological feature locations comprises:

selecting the first and second turning points as two of the template feature locations;

selecting one or more pre-QRS template feature locations each having a first timing relationship with the first turning point from the pre-QRS segment, the first timing relationship being a first function of the arrhythmic heart rate parameter; and

selecting one or more post-QRS template feature locations each having a second timing relationship with the second turning point from the post-QRS segment, the second timing relationship being a second function of the arrhythmic heart rate parameter.

20. The method of claim 18, wherein determining the morphological feature locations comprises:

selecting the first and second turning points as two of the template feature locations; and

selecting a plurality of QRS template feature locations from the QRS segment, the selected QRS template feature locations being a function of the arrhythmic heart rate parameter.

21. The method of claim 20, wherein selecting the plurality of QRS template feature locations comprises determining a quantity of the QRS template feature locations based on the arrhythmic heart rate parameter.

22. The method of claim 17, wherein determining the morphological feature locations comprises:

locating a template fiducial point on the template waveform and a corresponding arrhythmic fiducial point on the arrhythmic waveform;

measuring feature time intervals each between one template morphological feature of the plurality of template morphological features and the template fiducial point; and

determining the arrhythmic feature locations of the plurality of corresponding arrhythmic morphological features on the arrhythmic waveform using the arrhythmic fiducial point and the measured feature time intervals.

23. A system for classifying cardiac arrhythmias, the system comprising:

a template feature input to receive a plurality of template morphological features of a template waveform associated with a template heart beat of a known cardiac rhythm;

an arrhythmic feature input to receive a plurality of arrhythmic morphological features of an arrhythmic waveform associated with an arrhythmic heart beat of an arrhythmia episode, the arrhythmic morphological features each corresponding to one of the template morphological features;

an arrhythmic heart rate input to receive an arrhythmic heart rate parameter representative of an arrhythmic heart rate associated with the arrhythmic waveform;

a rate-dependent weighting module coupled to the template feature input, the arrhythmic feature input, and the arrhythmic heart rate input, the rate-dependent weighting module adapted to produce weighted template morphological features and weighted arrhythmic morphological features using at least the plurality of template morphological features, the plurality of arrhythmic morphological features, and the arrhythmic heart rate parameter; and

a correlation coefficient calculator, coupled to the rate-dependent weighting module, to calculate at least one correlation coefficient representative of a correlation between the weighted arrhythmic morphological features and the weighted template morphological features.

24. The system of claim 23, further comprising an arrhythmia classification module, coupled to the correlation coefficient calculator, to classify the arrhythmia episode based on the at least one correlation coefficient.

25. The system of claim 23, wherein the rate-dependent weighting module comprises a weighting factor generator to produce weighting factors each as a function of at least the arrhythmic heart rate parameter and is adapted to produce the weighted template morphological features and the weighted arrhythmic morphological features by applying one of the weighting factors to each of the template and arrhythmic morphological features.

26. The system of claim 25, further comprising a template heart rate input, coupled to the rate-dependent weighting module, to receive a template heart rate parameter representative of a template heart rate associated with the template waveform, and wherein the weighting factor generator is adapted to produce weighting factors each as a function of at least the arrhythmic heart rate parameter and the template heart rate parameter.

27. The system of claim 25, further comprising an arrhythmic heart rate estimator coupled to the arrhythmic heart rate input, the arrhythmic heart rate estimator adapted to produce the arrhythmic heart rate parameter by estimating an instantaneous heart rate associated with the arrhythmic heart beat.

28. The system of claim 27, wherein the arrhythmic heart rate estimator is further adapted to produce the arrhythmic heart rate parameter by mapping the estimated instantaneous heart rate to one heart rate range of a plurality of predetermined heart rate ranges.

29. The system of claim 28, wherein the weighting factor generator is adapted to map the arrhythmic heart rate parameter to one set of weight factors of a plurality of sets of predetermined weighting factors.

30. A method for classifying cardiac arrhythmias, the method comprising:

receiving a plurality of template morphological features of a template waveform associated with a template heart beat of a known cardiac rhythm;

receiving a plurality of arrhythmic morphological features of an arrhythmic waveform associated with an arrhythmic heart beat of an arrhythmia episode, the arrhythmic morphological features each corresponding to one of the template morphological features;

receiving an arrhythmic heart rate parameter representative of an arrhythmic heart rate associated with the arrhythmic waveform;

producing a plurality of weighting factors using at least the arrhythmic heart rate parameter;

producing weighted template morphological features and weighted arrhythmic morphological features by applying a weighting factor of the plurality of weighting factors to each of the template morphological features and each of the arrhythmic morphological features;

analyzing a correlation between the weighted template morphological features and the weighted arrhythmic morphological features; and

classifying the arrhythmia episode based on the correlation.

31. The method of claim 30, further comprising receiving a template heart rate parameter representative of a template heart rate associated with the template waveform, and wherein producing the plurality of weighting factors comprises producing the plurality of weighting factors using at least the arrhythmic heart rate parameter and the template heart rate parameter.

32. The method of claim 31, further comprising producing the arrhythmic heart rate parameter by estimating an instantaneous heart rate associated with the arrhythmic heart beat.

33. The method of claim 32, wherein producing the plurality of weighting factors comprises:

mapping the estimated instantaneous heart rate to one heart rate range of a plurality of predetermined heart rate ranges; and

mapping the heart rate range to one set of weight factors of a plurality of sets of predetermined weighting factors.

34. The method of claim 30, wherein producing the plurality of weighting factors comprises producing weighting factors to allow a relative weight of each of template QRS morphological features to increase with the arrhythmic heart rate, the template QRS morphological features each being one of the plurality of template morphology features extracted from a QRS complex of the template waveform.

35. The method of claim 34, wherein producing the plurality of weighting factors comprises producing weighting factors giving the template QRS morphological features relatively high weights and giving remaining template morphological features of the plurality of template morphological features relatively low weights.