SEQUENTIAL DISCRIMINATION APPROACH FOR DETECTING TREATABLE CARDIAC RHYTHMS

• US 20120095520A1
• Filed: 10/13/2010
• Published: 04/19/2012
• Est. Priority Date: 10/13/2010
• Status: Active Grant

• Current Assignee: Medtronic Incorporated (Medtronic Plc)
• Original Assignee: Medtronic Incorporated (Medtronic Plc)
• Inventors: Zhang, Xusheng, Stadler, Robert W., Gillberg, Jeffrey M.
• Granted Patent: US 8,170,654 B1
• Time in Patent Office: Days
• US Class Current: 607/15
• CPC Class Codes: A61B 5/287 Holders for multiple electrodes, e.g. electrode catheters for electrophysiological study [EPS]
• CPC Class Codes: A61B 5/361 Detecting fibrillation
• CPC Class Codes: A61B 5/363 Detecting tachycardia or bradycardia
• CPC Class Codes: A61B 5/686 Permanently implanted devices, e.g. pacemakers, other stimulators, biochips A61B5/6861 takes precedence
• CPC Class Codes: A61B 5/7239 using differentiation including higher order derivatives
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: A61N 1/3624 occurring in the atrium, i.e. atrial tachycardia
• CPC Class Codes: A61N 1/39622 Pacing therapy
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

A system and method for use in a medical device for discriminating cardiac events establishes population-based thresholds corresponding to cardiac signal morphology metrics for discriminating between a first cardiac event and a second cardiac event. A population-based threshold criterion for discriminating cardiac events is established. The population-based threshold criterion is applied to a cardiac signal segment and the segment is classified if the criterion is satisfied. A patient-specific threshold is established in response to the sensed cardiac signal segment not being classified after applying the population-based threshold criterion. The sensed signal segment is classified in response to the patient-specific threshold comparison.

Claims

1. A method of classifying cardiac signals according to a cardiac event type, comprising:

establishing a plurality of population-based thresholds corresponding to a plurality of cardiac signal morphology metrics for discriminating between a first cardiac event and a second cardiac event;

establishing a first threshold criterion for discriminating cardiac events, the first threshold criterion comprising a comparison between a cardiac signal morphology metric and one established population-based threshold of the established plurality of population-based thresholds;

sensing a cardiac signal segment in a patient;

applying the first threshold criterion to the sensed cardiac signal segment;

classifying the cardiac signal segment only in response to the applying the first threshold criterion being satisfied;

establishing a patient-specific threshold in response to the sensed cardiac signal segment not being classified after applying the first threshold criterion;

computing a metric of the sensed cardiac signal segment and comparing the metric to the established patient-specific threshold; and

classifying the sensed signal segment in response to the comparing of the metric to the established patient-specific threshold.

2. The method of claim 1, further comprising:

establishing a sequence of population-based threshold criteria;

advancing to a next population based threshold criterion in the sequence in response to the first threshold criterion not being satisfied;

classifying the sensed cardiac signal segment in response to the earliest criterion in the sequence being satisfied; and

cancelling a remaining population-based threshold criterion in the sequence in response to classifying the cardiac signal segment.

3. The method of claim 1, wherein establishing the plurality of population-based thresholds comprises:

sensing a cardiac signal for a plurality of time segments in a population of patients;

determining the plurality of morphology metrics for each of the plurality of time segments;

classifying each of the plurality of time segments according to a cardiac event;

plotting the plurality of morphology metrics for the classified time segments; and

defining a population-based threshold that substantially separates a first cluster of classified cardiac signal time segments from a second cluster of classified cardiac signal time segments.

4. The method of claim 1, wherein the plurality of cardiac signal morphology metrics comprises at least one of a low slope content, a normalized mean rectified amplitude, a spectral width, a signal overall variability, and an RR interval variability.

5. The method of claim 3, wherein establishing a plurality of population-based thresholds comprises:

plotting a multi-dimensional scatter plot of at least two of the plurality of morphology metrics for the plurality of classified cardiac signal time segments; and

defining a population-based threshold for one of the plurality of morphology metrics as a function of another of the plurality of morphology metrics based on the multi-dimensional plot, the threshold defined to separate the first cluster of classified cardiac signal time segments from the second cluster of classified cardiac signal time segments.

6. The method of claim 5, wherein the second cluster of classified cardiac signal segments overlaps a third cluster of classified cardiac signal segments, the established population-based threshold defining a one-way classification criterion for classifying only the first cluster of cardiac signal segments in response to the population-based threshold.

7. The method of claim 2, wherein establishing a sequence of population-based threshold criteria comprises determining the first threshold criterion as one of the population-based threshold criteria that results in a highest frequency of cardiac signal segment classifications.

8. The method of claim 7, wherein the highest frequency of cardiac signal segment classifications comprises a highest frequency of a treatable cardiac signal segment classification.

9. The method of claim 1, wherein establishing a patient-specific threshold comprises computing a characteristic of an earlier-occurring cardiac signal segment.

10. The method of claim 1, wherein establishing a patient-specific threshold comprises computing a characteristic of a currently occurring cardiac signal segment.

11. The method of claim 1, wherein establishing a patient-specific threshold comprises:

establishing a classification threshold for detecting a treatable rhythm onset;

classifying the sensed cardiac signal segment as treatable in response to the established classification threshold for detecting a treatable rhythm onset being detected; and

classifying the sensed cardiac signal segment as non-treatable in response to the established classification threshold for detecting a treatable rhythm onset not being detected.

12. The method of claim 1, wherein establishing a plurality of population-based thresholds comprises determining a correlation between a first one of the plurality of morphology metrics determined for a time segment and a second one of the plurality of morphology metrics determined for the same time segment.

13. A medical device system for classifying cardiac signals according to a cardiac event type, comprising:

a processor configured to

establish a plurality of population-based thresholds corresponding to a plurality of cardiac signal morphology metrics for discriminating between a first cardiac event and a second cardiac event, and

establish a first threshold criterion for discriminating cardiac events, the first threshold criterion comprising a comparison between a cardiac signal morphology metric and one established population-based threshold of the plurality of population-based thresholds;

a plurality of electrodes for sensing a cardiac signal segment;

a programmable memory storing the first population-based threshold criterion; and

a controller configured to;

apply the first threshold criterion to the sensed cardiac signal segment;

classify the sensed cardiac signal segment only in response to the applying the first threshold criterion being satisfied;

establish a patient-specific threshold in response to the sensed cardiac signal segment not being classified after applying the first threshold criterion;

compute a metric of the sensed cardiac signal segment and comparing the metric to the established patient-specific threshold; and

classify the sensed cardiac signal segment in response to the comparing of the metric to the established patient-specific threshold.

14. The system of claim 13, wherein the processor is further configured to establish a sequence of population-based threshold criteria;

and

the controller is configured to advance to a next population based threshold criterion in the sequence in response to the first threshold criterion not being satisfied, classify the sensed cardiac signal segment in response to the earliest criterion in the sequence being satisfied, and cancel a remaining population-based threshold criterion in the sequence in response to classifying the cardiac signal segment.

15. The system of claim 13, wherein establishing the plurality of population-based thresholds comprises:

sensing a cardiac signal for a plurality of time segments in a population of patients;

determining the plurality of morphology metrics for each of the plurality of time segments;

classifying each of the plurality of time segments according to a cardiac event;

plotting the plurality of morphology metrics for the classified time segments; and

defining a population-based threshold that substantially separates a first cluster of classified cardiac signal time segments from a second cluster of classified cardiac signal time segments.

16. The system of claim 13, wherein the plurality of cardiac signal morphology metrics comprises at least one of a low slope content, a normalized mean rectified amplitude, a spectral width, a signal overall variability, and an RR interval variability.

17. The system of claim 15, wherein establishing a plurality of population-based thresholds comprises

plotting a multi-dimensional scatter plot of at least two of the plurality of morphology metrics for the plurality of classified cardiac signal time segments;

and

defining a population-based threshold for one of plurality of morphology metrics as a function of another of the plurality of morphology metrics based on the multi-dimensional plot, the threshold defined to separate the first cluster of classified cardiac signal time segments from the second cluster of classified cardiac signal time segments.

18. The system of claim 17, wherein the second cluster of classified cardiac signal segments overlaps a third cluster of classified cardiac signal segments, the established population-based threshold defining a one-way classification criterion for classifying only the first cluster of cardiac signal segments in response to the population-based threshold.

19. The system of claim 14, wherein establishing a sequence of population-based threshold criteria comprises determining a first one of the population-based threshold criteria that results in a highest frequency of cardiac signal segment classifications.

20. The system of claim 19, wherein the highest frequency of cardiac signal segment classifications comprises a highest frequency of a treatable cardiac signal segment classification.

21. The system of claim 13, wherein establishing a patient-specific threshold comprises computing a characteristic of an earlier-occurring cardiac signal segment.

22. The system of claim 13, wherein establishing a patient-specific threshold comprises establishing a classification threshold for detecting a treatable rhythm onset, and wherein the controller is further configured to classify the sensed cardiac signal segment as treatable in response to the established classification threshold for detecting a treatable rhythm onset being detected, and classify the sensed cardiac signal segment as non-treatable in response to the established classification threshold for detecting a treatable rhythm onset not being detected.

23. The system of claim 13, wherein establishing a plurality of population-based thresholds comprises determining a correlation between a first one of the plurality of morphology metrics determined for a time segment and a second one of the plurality of morphology metrics determined for the same time segment.

24. The system of claim 13, wherein establishing a patient-specific threshold comprises computing a characteristic of a currently occurring cardiac signal segment.

25. A computer-readable medium storing a set of instructions which cause a processor of a medical device system to:

establish a plurality of population-based thresholds corresponding to a plurality of cardiac signal morphology metrics for discriminating between a first cardiac event and a second cardiac event;

establish a first threshold criterion for discriminating cardiac events, the first threshold criterion comprising a comparison between a cardiac signal morphology metric and one established population-based threshold of the established plurality of population-based thresholds;

sense a cardiac signal segment;

apply the first threshold criterion to the sensed cardiac signal segment;

classify the sensed cardiac signal segment only in response to the applying the first threshold criterion being satisfied;

establish a patient-specific threshold in response to the sensed cardiac signal segment not being classified after applying the first threshold criterion;

compute a metric of the sensed cardiac signal segment and comparing the metric to the established patient-specific threshold; and

classify the sensed signal segment in response to the comparing of the metric to the established patient-specific threshold.