Method and apparatus for morphology-based arrhythmia classification using cardiac and other physiological signals

• US 20070149890A1
• Filed: 12/22/2005
• Published: 06/28/2007
• Est. Priority Date: 12/22/2005
• Status: Active Grant

• Current Assignee: Cardiac Pacemakers Incorporated (Boston Scientific Corporation)
• Original Assignee: Cardiac Pacemakers Incorporated (Boston Scientific Corporation)
• Inventors: Kim, Jaeho, Li, Dan, Lee, Kent, Cazares, Shelley
• Granted Patent: US 7,582,061 B2
• Time in Patent Office: Days
• US Class Current: 600/515
• 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: A61N 1/3621 for treating or preventing abnormally high heart rate
• CPC Class Codes: A61N 1/36521 the parameter being derived from measurement of an electrical impedance
• CPC Class Codes: A61N 1/36542 controlled by body motion, e.g. acceleration
• CPC Class Codes: A61N 1/36564 controlled by blood pressure
• CPC Class Codes: A61N 1/36571 controlled by blood flow rate, e.g. blood velocity or cardiac output
• CPC Class Codes: A61N 1/36578 controlled by mechanical motion of the heart wall, e.g. measured by an accelerometer or microphone
• CPC Class Codes: A61N 1/36585 controlled by two or more physical parameters
• CPC Class Codes: A61N 1/3925 Monitoring; Protecting
• CPC Class Codes: A61N 1/39622 Pacing therapy
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A tachyarrhythmia detection and classification system classifies tachyarrhythmias based on an analysis of morphological features of a cardiac signal enhanced by using one or more physiological parameters indicative of hemodynamic stability and/or activity level. The tachyarrhythmia detection and classification system computes a measure of similarity between an arrhythmic waveform of the cardiac signal a template waveform for that cardiac signal, such as a correlation coefficient representative of the correlation between morphological features of the arrhythmic waveform and morphological features of the template waveform. A detected tachyarrhythmia episode is classified by comparing the measure of similarity to a threshold that is dynamically adjusted using the one or more physiological parameters.

Claims

1. A tachyarrhythmia detection and classification system, comprising:

a cardiac sensing circuit to sense at least one cardiac signal;

a tachyarrhythmia detector, coupled to the cardiac sensing circuit, to detect a tachyarrhythmia episode from the at least one cardiac signal;

one or more physiological sensors to sense one or more physiological signals;

a physiological parameter generator, coupled to the one or more physiological sensors, to derive one or more physiological parameters from the one or more physiological signals; and

a morphology-based tachyarrhythmia classification module coupled to the tachyarrhythmia detector, the morphology-based tachyarrhythmia classification module including;

a feature extractor, coupled to the cardiac sensing circuit, to extract arrhythmic morphological features of an arrhythmic heart beat of the detected tachyarrhythmia episode from the at least one cardiac signal;

a similarity analyzer, coupled to the cardiac sensing circuit, to compute a measure of similarity between the arrhythmic morphological features and template morphological features of a template heart beat of a known type cardiac rhythm;

a classifier, coupled to the similarity analyzer, to classify the arrhythmic heart beat using the measure of similarity and a dynamically adjustable beat classification threshold (T); and

a threshold generator, coupled to the physiological parameter generator and the classifier, to adjust the beat classification threshold (T) using the one or more physiological parameters.

2. The system of claim 1, wherein the similarity analyzer comprises a correlation analyzer adapted to compute a correlation coefficient between the arrhythmic morphological features and the template morphological features for each arrhythmic heart beat of a plurality of arrhythmic heart beats of the detected tachyarrhythmia episode, and the classifier comprises:

a beat classifier including a beat comparator having a first input receiving the correlation coefficient for each of the plurality of arrhythmic heart beats, a second input receiving the beat classification threshold (T), and an output indicating a classification for the each of the plurality of arrhythmic heart beats as one of a first type tachyarrhythmic heart beat and a second type tachyarrhythmic heart beat;

a beat counter, coupled to the beat classifier, to count a number of the arrhythmic heart beats classified as the first type tachyarrhythmic heart beat; and

an episode classifier including an episode comparator having a first input to receive the number of the arrhythmic heart beats classified as the first type tachyarrhythmic heart beat, a second input receiving an episode classification threshold, and an output indicating a classification of the detected tachyarrhythmia episode as one of a first type tachyarrhythmia and a second type tachyarrhythmia.

3. The system of claim 1, wherein the threshold generator comprises a mapping module to map the one or more physiological parameters to one of a plurality of predetermined values of the beat classification threshold (T).

4. The system of claim 3, wherein the mapping module comprises a look-up table relating the one or more physiological parameters to the plurality of predetermined values of the beat classification threshold (T).

5. The system of claim 1, wherein the one or more physiological signals comprise at least one hemodynamic signal indicative of hemodynamic performance, the one or more physiological parameters comprise a hemodynamic stability parameter indicative of a level of hemodynamic stability, the one or more physiological sensors comprise a hemodynamic sensor to sense the hemodynamic signal, the physiological parameter generator is adapted to derive the hemodynamic stability parameter (S) from the hemodynamic signal, and the threshold generator is adapted to adjust the beat classification threshold (T) using the hemodynamic stability parameter (S).

6. The system of claim 5, wherein the physiological parameter generator comprises:

a template parameter generator to produce a template hemodynamic parameter (H_T);

an arrhythmic parameter generator to measure an arrhythmic hemodynamic parameter (H_A) from the at least one hemodynamic signal sensed during the detected tachyarrhythmia episode; and

a hemodynamic stability analyzer to produce the hemodynamic stability parameter (S) as a function of the template hemodynamic parameter (H_T) and the arrhythmic hemodynamic parameter (H_A).

7. The system of claim 6, wherein the at least one hemodynamic signal comprises an impedance signal indicative of one of a transthoracic impedance and a cardiac impedance, the hemodynamic sensor comprises an impedance sensor to sense the impedance signal, and the physiological parameter generator is adapted to derive the hemodynamic stability parameter (S) from the impedance signal.

8. The system of claim 7, wherein the physiological parameter generator comprises a stroke impedance detector adapted to measure a stroke impedance (Δ

Z) from the impedance signal, the stroke impedance being a difference between a maximum impedance (Z_max) and a minimum impedance (Z_min) over a cardiac cycle, and wherein the template hemodynamic parameter (H_T) is a template stroke impedance (Δ

Z_T), and the arrhythmic hemodynamic parameter (H_A) is an arrhythmic stroke impedance (Δ

Z_A).

9. The system of claim 6, wherein the at least one hemodynamic signal comprises a pressure signal indicative of an arterial pressure, the hemodynamic sensor comprises a pressure sensor to sense the pressure signal, and the physiological parameter generator to derive the hemodynamic stability parameter (S) from the pressure signal.

10. The system of claim 9, wherein the pressure signal comprises a pulmonary artery pressure (PAP) signal indicative of PAP, and the pressure sensor comprises an implantable PAP sensor.

11. The system of claim 9, further comprising a rate detector, coupled to the cardiac sensing circuit, to detect a heart rate from the at least one cardiac signal, and wherein the physiological parameter generator comprises a pressure detector to measure a pressure (P) from the pressure signal, wherein the template hemodynamic parameter (H_T) is a template pressure (P_T) being a pressure value estimated for an arrhythmic heart rate detected during the detected tachyarrhythmia episode, and the arrhythmic hemodynamic parameter (H_A) is an arrhythmic pressure (P_A), and wherein the template parameter generator is adapted to produce the template pressure (P_T) as a given function of the arrhythmic heart rate.

12. The system of claim 5, wherein the at least one hemodynamic signal comprises one of a signal indicative of blood flow, a signal indicative of heart sounds, an oximetry signal indicative of blood oxygenation, a plethysmographic signal indicative of cardiac output, a respiratory signal, and a cardiac contractility signal, the hemodynamic sensor comprises on of a flow sensor to sense the signal indicative of blood flow, a heart sound sensor to sense the signal indicative of heart sounds, a pulse oximetry sensor to sense the oximetry signal, a photoplethysmography sensor to sense the plethysmographic signal, a respiratory sensor to sense the respiratory signal, and a cardiac contractility sensor to sense the cardiac contractility signal, and the physiological parameter generator to derive the hemodynamic stability parameter (S) from one of the signal indicative of blood flow, the signal indicative of heart sounds, the oximetry signal, the plethysmographic signal, the respiratory signal and the cardiac contractility signal.

13. The system of claim 5, wherein the one or more physiological signals comprise first and second hemodynamic signals each indicative of hemodynamic performance, the one or more physiological parameters comprise a hemodynamic stability parameter indicative of a level of hemodynamic stability, the one or more physiological sensors comprise a first hemodynamic sensor to sense the first hemodynamic signal and a second hemodynamic sensor to sense the second hemodynamic signal, the physiological parameter generator is adapted to derive the hemodynamic stability parameter (S) from the first and second hemodynamic signals, and the threshold generator is adapted to adjust the beat classification threshold (T) using the hemodynamic stability parameter (S).

14. The system of claim 13, wherein the physiological parameter generator comprises:

a template parameter generator to produce a first template hemodynamic parameter (H_1T) for the first hemodynamic signal, a second template hemodynamic parameter (H_2T) for the second hemodynamic signal, and a template parameter vector U_T=[H_1T, H_2T] over a predetermined number of heart beats;

an arrhythmic parameter generator to measure a first arrhythmic hemodynamic parameter (H_1A) from the first hemodynamic signal sensed during the detected tachyarrhythmia episode, a second arrhythmic hemodynamic parameter (H_2A) from the second hemodynamic signal sensed during the detected tachyarrhythmia episode, and an arrhythmic parameter vector U_A=[H_1A, H_2A]; and

a hemodynamic stability analyzer to produce the hemodynamic stability parameter (S) as a function of the template parameter vector (U_T) and the arrhythmic parameter vector (U_A).

15. The system of claim 14, wherein the hemodynamic stability analyzer comprises a Mahalanobis distance computer adapted to compute a Mahalanobis distance from the arrhythmic parameter vector (U_A) to the template parameter vector (U_T) and produce the hemodynamic stability parameter (S) as a function of the Mahalanobis distance.

16. The system of claim 13, wherein the first hemodynamic signal is an impedance signal indicative one of a transthoracic impedance and a cardiac impedance, the second hemodynamic signal is a pressure signal indicative of an arterial pressure, the first hemodynamic sensor is an impedance sensor to sense the impedance signal, the second hemodynamic sensor is a pressure sensor to sense the pressure signal, and the physiological parameter generator to derive the hemodynamic stability parameter (S) from the impedance signal and the pressure signal.

17. The system of claim 5, wherein the one or more physiological signals comprise the at least one hemodynamic signal and an activity signal indicative of an activity level, the one or more physiological parameters comprise a hemodynamic stability parameter (S) indicative of a level of hemodynamic stability and an activity parameter (A) indicative of the activity level, the one or more physiological sensors comprise at least one hemodynamic sensor to sense the at least one hemodynamic signal and an activity sensor to sense the activity signal, the physiological parameter generator is adapted to derive the hemodynamic stability parameter (S) from the at least one hemodynamic signal and to derive the activity parameter (A) from the activity signal, and the threshold generator is adapted to adjust the beat classification threshold (T) using the hemodynamic stability parameter (S) and the activity parameter (A).

18. The system of claim 17, wherein the threshold generator comprises a threshold update computer adapted to determine a change in the beat classification threshold related to hemodynamic stability (Δ

T_S) in response to a change in the hemodynamic stability parameter (Δ

S), to determine a change in the beat classification threshold related to activity (Δ

T_A) in response to a change in the activity parameter (Δ

A), and to compute a change in the beat classification threshold (Δ

T) as a function of the change in the hemodynamic stability related beat classification threshold (Δ

T_S) and the change in the activity related beat classification threshold (Δ

T_A), and wherein the threshold generator is adapted to update the beat classification threshold (T) by the change in the beat classification threshold (Δ

T).

19. The system of claim 1, further comprising a rate detector, coupled to the cardiac sensing circuit, to detect a heart rate from the at least one cardiac signal, and wherein the threshold generator comprises a normal heart rate estimator adapted to produce an estimated normal heart rate range using the one or more physiological signals, a normal heart rate associated with the known type cardiac rhythm, and demographic information, the threshold generator adopted to adjust the beat classification threshold (T) using an arrhythmic heart rate detected during the detected tachyarrhythmia episode and the estimated normal heart rate range.

20. The system of claim 19, wherein the one or more physiological signals comprise at least one activity signal indicative of an activity level, and the one or more physiological sensors include at least one activity sensor to sense the activity signal.

21. A method for detecting and classifying tachyarrhythmias, the method comprising:

sensing at least one cardiac signal;

detecting a tachyarrhythmia episode from the at least one cardiac signal;

extracting arrhythmic morphological features of an arrhythmic heart beat of the detected tachyarrhythmia episode from the at least one cardiac signal;

computing a measure of similarity between the arrhythmic morphological features and template morphological features of a template heart beat of a known type cardiac rhythm for each arrhythmic heart beat of a plurality of arrhythmic heart beats;

classifying the detected tachyarrhythmia episode using values of the measure of similarity computed for the plurality of arrhythmic heart beats and a dynamically adjustable beat classification threshold (T);

sensing one or more physiological signals;

deriving one or more physiological parameters from the one or more physiological signals; and

adjusting the beat classification threshold (T) using the one or more physiological parameters.

22. The method of claim 21, wherein the measure of similarity comprises a correlation coefficient, and wherein classifying the detected tachyarrhythmia episode comprises:

comparing the correlation coefficient for each of the predetermined number of arrhythmic heart beats to the beat classification threshold (T),

indicating a classification for the each of the predetermined number of arrhythmic heart beats as one of a first type tachyarrhythmic heart beat and a second type tachyarrhythmic heart beat;

counting a number of the arrhythmic heart beats classified as the first type tachyarrhythmic heart beat;

comparing the number of the arrhythmic heart beats classified as the first type tachyarrhythmic heart beat to an episode classification threshold; and

indicating a classification of the detected tachyarrhythmia episode as a first type tachyarrhythmia if the number of the arrhythmic heart beats classified as the first type tachyarrhythmic heart beat is equal to or greater than the episode classification threshold and as a second type tachyarrhythmic heart beat if the number of the arrhythmic heart beats classified as the first type tachyarrhythmic heart beat is below the episode classification threshold.

23. The method of claim 21, wherein adjusting the beat classification threshold (T) comprises mapping the one or more physiological parameters to one of a plurality of predetermined values of the beat classification threshold (T).

24. The method of claim 21, wherein sensing the one or more physiological signals comprises sensing at least one hemodynamic signal indicative of hemodynamic performance, deriving the one or more physiological parameters comprises deriving a hemodynamic stability parameter (S) from the at least one hemodynamic signal, and adjusting the beat classification threshold (T) comprises using adjusting the beat classification threshold (T) using the hemodynamic stability parameter (S).

25. The method of claim 24, wherein deriving the hemodynamic stability parameter (S) from the at least one hemodynamic signal comprises:

producing a template hemodynamic parameter (H_T);

measuring an arrhythmic hemodynamic parameter (H_A) from the hemodynamic signal sensed during the detected tachyarrhythmia episode; and

producing the hemodynamic stability parameter (S) as a function of the template hemodynamic parameter (H_T) and the arrhythmic hemodynamic parameter (H_A).

26. The method of claim 25, wherein producing the hemodynamic stability parameter (S) comprises:

computing a change (δ

) being a ratio of a difference between the arrhythmic hemodynamic parameter (H_A) and the template hemodynamic parameter (H_T) to the template hemodynamic parameter (H_T); and

producing the hemodynamic stability parameter (S) as a function of the change (δ

).

27. The method of claim 25, wherein sensing the at least one hemodynamic signal comprises sensing an impedance signal indicative of one of a transthoracic impedance and a cardiac impedance, and deriving the hemodynamic stability parameter (S) comprise deriving the hemodynamic stability parameter (S) from the impedance signal.

28. The method of claim 27, wherein deriving the hemodynamic stability parameter (S) from the impedance signal comprises measuring a stroke impedance (Δ

Z) from the impedance signal, the stroke impedance being a difference between a maximum impedance (Z_max) and a minimum impedance (Z_min) over a cardiac cycle, and wherein the template hemodynamic parameter (H_T) is a template stroke impedance (Δ

Z_T), and the arrhythmic hemodynamic parameter (H_A) is an arrhythmic stroke impedance (Δ

Z_A).

29. The method of claim 25, wherein sensing the at least one hemodynamic signal comprises sensing a pressure signal indicative of an arterial pressure, and deriving the hemodynamic stability parameter (S) comprise deriving the hemodynamic stability parameter (S) from the pressure signal.

30. The method of claim 29, further comprising detecting a heart rate from the at least one cardiac signal, and wherein deriving the hemodynamic stability parameter (S) comprises:

measuring a pressure (P) from the pressure signal;

producing a template pressure (P_T) as a given function of an arrhythmic heart rate detected during the detected tachyarrhythmia episode; and

measuring an arrhythmic pressure (P_A) from the pressure signal sensed during the detected tachyarrhythmic episode,

and wherein the template hemodynamic parameter (H_T) is the template pressure (P_T), and the arrhythmic hemodynamic parameter (H_A) is the arrhythmic pressure (P_A).

31. The method of claim 25, wherein the at least one hemodynamic signal comprises one of a signal indicative of blood flow, a signal indicative of heart sounds, an oximetry signal indicative of blood oxygenation, a plethysmographic signal indicative of cardiac output, a respiratory signal, and a cardiac contractility signal.

32. The method of claim 24, wherein sensing the one or more physiological signals comprises sensing first and second hemodynamic signals each indicative of hemodynamic performance, deriving the one or more physiological parameters comprises deriving a hemodynamic stability parameter (S) from the first and second hemodynamic signals, and adjusting the beat classification threshold (T) comprises adjusting the beat classification threshold (T) using the hemodynamic stability parameter (S).

33. The method of claim 32, wherein deriving the hemodynamic stability parameter (S) comprises:

producing a first template hemodynamic parameter (H_1T) for the first hemodynamic signal, a second template hemodynamic parameter (H_2T) for the second hemodynamic signal, and a template parameter vector U_T=[H_1T, H_2T] over a predetermined number of heart beats of the know type cardiac rhythm;

measuring a first arrhythmic hemodynamic parameter (H_1A) from the first hemodynamic signal sensed during the detected tachyarrhythmia episode, a second arrhythmic hemodynamic parameter (H_2A) from the second hemodynamic signal sensed during the detected tachyarrhythmia episode, and an arrhythmic parameter vector U_A=[H_1A, H_2A], and

producing the hemodynamic stability parameter (S) as a function of the template parameter vector (U_T) and the arrhythmic parameter vector (U_A).

34. The method of claim 33, wherein producing the hemodynamic stability parameter (S) comprises:

computing a Mahalanobis distance from the arrhythmic parameter vector (U_A) to the template parameter vector (U_T); and

producing the hemodynamic stability parameter (S) as a function of the Mahalanobis distance.

35. The method of claim 32, wherein the first hemodynamic signal is an impedance signal indicative of one of a transthoracic impedance and a cardiac impedance, the second hemodynamic signal is a pressure signal indicative of an arterial pressure, and deriving the hemodynamic stability parameter (S) comprises deriving the hemodynamic stability parameter (S) from the impedance signal and the pressure signal.

36. The method of claim 24, wherein sensing the one or more physiological signals comprises sensing the at least one hemodynamic signal and an activity signal indicative of an activity level, deriving the one or more physiological parameters comprises deriving the hemodynamic stability parameter (S) from the at least one hemodynamic signal and an activity parameter (A) indicative of the activity level from the activity signal, and adjusting the beat classification threshold (T) comprises adjusting the beat classification threshold (T) using the hemodynamic stability parameter (S) and the activity parameter (A).

37. The method of claim 36, wherein adjusting the beat classification threshold (T) comprises:

determining a change in the beat classification threshold related to hemodynamic stability (Δ

T_S) in response to a change in the hemodynamic stability parameter (Δ

S);

determining a change in the beat classification threshold related to activity (Δ

T_A) in response to a change in the activity parameter (Δ

A); and

updating the beat classification threshold (T) by a change in the beat classification threshold (Δ

T) being a function of the change in the hemodynamic stability related beat classification threshold (Δ

T_S) and the change in the activity related beat classification threshold (Δ

T_A).

38. The method of claim 21, further comprising detecting a heart rate from the at least one cardiac signal, and wherein adjusting the beat classification threshold (T) comprises:

producing an estimated normal heart rate range using the one or more physiological signals, a normal heart rate associated with the known type cardiac rhythm, and demographic information; and

adjusting the beat classification threshold (T) using an arrhythmic heart rate detected during the detected tachyarrhythmia episode and the estimated normal heart rate range.

39. The method of claim 38, wherein the one or more physiological signals comprises at least one activity signal indicative of an activity level.

40. The method of claim 39, wherein adjusting the beat classification threshold (T) comprises setting the beat classification threshold (T) to a first predetermined value if the arrhythmic heart rate falls within the estimated normal heart rate range and a second value if the arrhythmic heart rate does not fall within estimated normal heart rate range.