Method and apparatus for ectopic beat detection

• US 8,315,694 B2
• Filed: 10/16/2008
• Issued: 11/20/2012
• Est. Priority Date: 10/16/2008
• Status: Expired due to Fees

• Current Assignee: Biotronik SE & Company KG
• Original Assignee: Biotronik CRM Patent AG (Biotronik SE & Company KG)
• Inventors: Lian, Jie, Muessig, Dirk
• Primary Examiner(s): Patel, Niketa
• Assistant Examiner(s): Edwards, Philip
• Application Number:
  US12/252,529
• Publication Number:
  US 20100099995A1
• Time in Patent Office: 1,496 Days
• Field of Search: 600/509
• US Class Current: 600/509
• CPC Class Codes: A61N 1/3621 for treating or preventing abnormally high heart rate
• CPC Class Codes: A61N 1/3702 Physiological parameters A61N1/365 takes precedence; evoked response A61N1/371

Abstract

Heart monitor for detecting ectopic beats in an input electrocardiogram signal that includes an electrocardiogram signal input and a morphological signal analyzer connected to the electrocardiogram signal input, the analyzer being adapted to generate a first time series of values representing the input electrocardiogram signal, a second signal analyzer adapted to generate generating a modified time series of values representing a trend of values of the first time series and a comparison stage being adapted to compare the first time series with the modified time series to thus detect ectopic beats. The invention further relates to a method for detecting ectopic beats in an input electrocardiogram signal that includes obtaining an electrocardiogram signal, generating from the electrocardiogram signal a first time series of values representing the input electrocardiogram signal, generating a modified time series of values representing a trend of values of the first time series and comparing the first time series with the modified time series to thus detect ectopic beats.

Claims

1. A heart monitor for detecting ectopic beats in an electrocardiogram signal comprising:

an implantable device configured to monitor an electrocardiogram signal subcutaneously;

an electrocardiogram signal input coupled with said implantable device and configured to obtain said electrocardiogram signal;

a first signal analyzer connected to said electrocardiogram signal input, said first signal analyzer configured to generate a first time series of values that represent each duration of a respective measured RR interval derived from said electrocardiogram signal;

a second signal analyzer configured to generate a modified time series of values that represent a trend of values of said first time series through application of one or more morphological operators comprising an erosion operator or a dilation operator or any combination thereof and to remove abnormal RR intervals from said first time series of values that represent each duration of the respective measured RR-interval, to thus derive the modified time series of RR-intervals that represent a trend of RR intervals; and

,

a comparison stage configured to compare said first time series with said modified time series and thus

detect ectopic beats from said modified time series generated with a first discrete kernel function of a first length or

detect non-sustained ventricular tachycardia from said modified time series generated with a second discrete kernel function of a second length wherein said second length is greater than said first length or

detect both ectopic beats from said modified time series generated with said first discrete kernel function of said first length and non-sustained ventricular tachycardia from said modified time series generated with said second discrete kernel function of said second length wherein said second length is greater than said first length.

2. The heart monitor of claim 1,

wherein said first signal analyzer is further configured to generate a QRS metrics time series of values that represent QRS metrics, wherein each value of said time series represents at least one of the QRS metrics, comprising

a width of a QRS complex, or

a positive or negative peak amplitude of the QRS complex, or

an absolute area under the QRS complex, or

maximum positive or negative slopes of the QRS complex, or

a dominant frequency component of the QRS complex, or

complexity measures of the QRS complex;

and,

wherein said second signal analyzer is configured to further apply said one or more morphological operators comprising said erosion operator or said dilation operator or said any combination thereof to the QRS metrics time series of values that represent said QRS metrics and to remove abnormal QRS metrics from said QRS metrics time series of values that represent said QRS metrics to thus derive another modified time series that represent a trend of said QRS metrics and use said trend of QRS metrics in combination with said modified time series in said comparison stage if at least one or both of said trend of QRS metrics and said modified time series classify a beat as abnormal.

3. The heart monitor of claim 2 wherein said comparison stage is configured to subtract said first time series from said modified time series and to thus generate difference signal values and detect said abnormal RR intervals or the QRS metrics through comparison of said difference signal values to at least one threshold value.

4. The heart monitor of claim 2 wherein said comparison stage is configured to generate a series of ratio values through division of each value of said first time series by an associated value of said modified time series and to compare each ratio value thus derived with at least one threshold value.

5. The heart monitor of claim 3 wherein said comparison stage is configured to generate both the difference signal values and configured to generate a series of ratio values through division of each value of said first time series by an associated value of said modified time series and to compare each difference signal value to at least one threshold value and to compare each ratio value with another at least one threshold value.

6. The heart monitor of claim 1 wherein said the second signal analyzer is configured to generate said modified time series through application of both an opening operator and a closing operator to said first time series to thus obtain said modified time series of values that represent the trend of values of said first time series.

7. The heart monitor of claim 6, wherein said the second signal analyzer is configured to generate said modified time series by applying said erosion operator followed by said dilation operator that together form the opening operator to suppress peaks in the first time series.

8. The heart monitor of claim 6 wherein the second signal analyzer is configured to generate said modified time series by applying said dilation operator followed by said erosion operator that together form the closing operator to suppress pits in the first time series.

9. The heart monitor of claim 1 further comprising a third signal analyzer that is configured to detect multiple cycles of consecutive short RR intervals or multiple cycles of consecutive long RR intervals or both to detect said non-sustained ventricular tachycardia.

10. The heart monitor of claim 1 wherein said first discrete kernel function of said first length is filled with zeros or wherein said second discrete kernel function of said second length is filled with zeros or wherein said first and said second discrete kernel functions are both filled with zeros to avoid a subtraction operation.

11. A method for detecting ectopic beats in an electrocardiogram signal comprising:

obtaining an electrocardiogram signal with an implantable device that is monitoring an electrocardiogram signal subcutaneously with an electrocardiogram signal input;

generating from said electrocardiogram signal a first time series of values that represent each duration of a respective measured RR interval derived from said input electrocardiogram signal via a first signal analyzer;

generating a modified time series of values that represent a trend of values of said first time series by applying one or more morphological operators comprising an erosion operator or a dilation operator or any combination thereof and to remove abnormal RR intervals from said first time series of values that represent each duration of the respective measured RR-interval, to thus derive the modified time series of RR-intervals that represent a trend of RR intervals via a second signal analyzer; and

,

comparing said first time series with said modified time series via a comparison stage and thus

detecting ectopic beats from said modified time series generated with a first discrete kernel function of a first length or

detecting non-sustained ventricular tachycardia from said modified time series generated with a second discrete kernel function of a second length wherein said second length is greater than said first length or

detecting both ectopic beats from said modified time series generated with said first discrete kernel function of said first length and non-sustained ventricular tachycardia from said modified time series generated with said second discrete kernel function of said second length wherein said second length is greater than said first length.

12. The method of claim 11, further comprising

generating from said electrocardiogram signal a QRS metrics time series comprises generating a time series of values that represent QRS metrics from said first time series, wherein each value of said time series represents at least one of the QRS metrics, comprising

a width of a QRS complex, or

a positive or negative peak amplitude of the QRS complex, or

an absolute area under the QRS complex, or

maximum positive or negative slopes of the QRS complex, or

a dominant frequency component of the QRS complex, or

complexity measures of the QRS complex;

and,

wherein said generating said modified time series comprises further applying said one or more morphological operators comprising said erosion operator or said dilation operator or said any combination thereof to the QRS metrics time series of values that represent said QRS metrics and removing abnormal QRS metrics from said first time series of values that represent QRS metrics and thus deriving another modified time series that represent a trend of said QRS metrics and using said trend of QRS metrics in combination with said modified time series in said comparison stage if at least one or both of said trend of QRS metrics and said modified time series classify a beat as abnormal.

13. The method of claim 12 wherein said comparing said first time series with said modified time series comprises subtracting said first time series from said modified time series to thus generate difference signal values and detecting abnormal RR intervals or QRS metrics by comparing said difference signal values to at least one threshold value.

14. The method of claim 12 wherein said comparing said first time series with said modified time series comprises generating a series of ratio values by dividing each value of said first time series by an associated value of said modified time series and comparing each ratio value thus derived with at least one threshold value.

15. The method of claim 13 wherein said comparing said first time series with said modified time series comprises generating both said difference signal values and wherein said comparing said first time series with said modified time series further comprises generating a series of ratio values by dividing each value of said first time series by an associated value of said modified time series and comparing each difference signal value to at least one threshold value and comparing each ratio value with another at least one threshold value.

16. The method of claim 11 wherein said step of generating said modified time series comprises applying both, an opening operator and a closing operator to said first time series to thus obtain said modified time series of values that represent the trend of values of said first time series.

17. The method of claim 16, wherein said generating said modified time series comprises applying said erosion operator followed by said dilation operator that together form the opening operator to suppress peaks in said first time series and/or wherein said generating said modified time series comprises applying said dilation operator followed by said erosion operator that together form the closing operator to suppress pits in the first time series.

18. The method of claim 17 further comprising processing the trend of values of the first time series to detect multiple cycles of consecutive short RR intervals or multiple cycles of consecutive long RR intervals or both and detecting said non-sustained ventricular tachycardia.

19. The method of claim 11 further comprising using said first discrete kernel function of said first length filled with zeros or using said second discrete kernel function of said second length filled with zeros or using said first and said second discrete kernel functions both filled with zeros to avoid a subtracting operation.