Method of classifying heart rhythms by analyzing several morphology defining metrics derived for a patient's QRS complex

• US 5,400,795 A
• Filed: 10/22/1993
• Issued: 03/28/1995
• Est. Priority Date: 10/22/1993
• Status: Expired due to Term

• Current Assignee: Pacesetter Incorporated (Abbott Laboratories)
• Original Assignee: Telectronics Pacing Systems, Inc (Abbott Laboratories)
• Inventors: Murphy, Anthony J., Bassin, David, Wickham, John
• Primary Examiner(s): Kamm, William E.
• Application Number:
  US08/141,805
• Time in Patent Office: 522 Days
• Field of Search: 728/702, 728/703, 728/704, 728/705
• US Class Current: 600/515
• CPC Class Codes: A61B 5/35 by template matching
• 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 1/3956 Implantable devices for applying electric shocks to the heart, e.g. for cardioversion
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: Y10S 128/925 Neural network

Abstract

A method, useable in implantable cardioverter defibrillators, for classifying heart rhythms of a patient by means of electrogram morphology. A number of features (metrics) of the electrogram are measured to form a description of the shape of individual electrograms, and the metrics are then converted into a cardiac rhythm diagnosis by means of a pattern classification technique.

Claims

1. A method of classifying heart rhythms of a patient by means of electrogram morphology, comprising the steps of:

storing a first set of at least two morphology-defining metrics derived from a QRS complex of a normal electrogram, said electrogram having a width and an amplitude, said metrics defining the shape of said electrogram including at least one of said width and amplitude;

establishing bounds of normal variation for said metrics which are combined to define a decision surface that includes first and second regions such that said first region represents normal heartbeats and said second region represents abnormal heartbeats;

providing an electrogram of a patient'"'"'s heart rhythm;

measuring a second set of metrics for a given heartbeat from said patient'"'"'s electrogram, said metrics of said second set corresponding to said metrics of said first set;

comparing the metrics of said second set with the metrics of said first set to decide which of said regions said given heartbeat falls into and to classify said given heartbeat as either normal or abnormal based on said decision;

repeating said measuring and comparing steps for a plurality of times; and

,

if a number of heartbeats are found to be abnormal, diagnosing the patient'"'"'s heart rhythm as ventricular arrhythmia.

2. A method according to claim 1, wherein said storing step includes a sub-step of storing a first set of at least three metrics selected from a group of metrics that includes:

a MOD metric, which comprises an absolute area under a QRS complex of a normal electrogram;

a PK metric, which comprises a value of a largest excursion of said QRS complex from its baseline, said value having the same polarity as said excursion;

a PVE metric, which comprises an amplitude of a largest positive deflection of said QRS complex;

a NVE metric, which comprises an amplitude of a largest negative deflection of said QRS complex;

a PKP metric, which comprises a peak-to-peak amplitude of said QRS complex;

a DIF metric, which comprises the difference between the absolute amplitudes of the largest positive deflection of said QRS complex and the largest negative deflection of said QRS complex;

a FIL metric, which comprises a peak-to-peak amplitude of said QRS complex after said complex has passed through a bandpass filter having a center frequency of about 25 Hz;

a SSA metric, which comprises a signed-square area metric in which a plurality of samples taken of said QRS complex are squared, their signs are restored, the signed squares are summed, the square root of the absolute value of the sum is taken, and the sign is restored thereto;

a QRS metric, which comprises a QRS angle Θ

calculated from the amplitude and temporal order of the largest positive and largest negative peaks of the QRS complex, as follows;

space="preserve" listing-type="equation">r=abs (smaller peak)÷

(larger peak),

space="preserve" listing-type="equation">Φ

=arcos (r),

and

if the first peak is positive and larger, Θ

=Φ

,

if the first peak is positive and smaller, Θ

=-Φ

,

if the first peak is negative and larger, Θ

=Φ

-180, and,

if the first peak is negative and smaller, Θ

=180-Φ

; and

, an ORD metric, which is defined as follows;

if the positive peak is first, ORD=smaller peak÷

larger peak, and

if the negative peak is first, ORD=-smaller peak÷

larger peak.

3. A method according to claim 2, wherein said storing step includes a sub-step of storing a first set of at least three metrics that includes said MOD metric, said PKP metric and said NVE metric period.

4. A method according to claim 2, wherein, if one or more of the MOD, PVE, NVE, DIF, FIL and SSA metrics are selected from said group of metrics, said storing step includes a further sub-step of amplitude normalizing said one or more metrics by dividing the value thereof by the value of said PKP metric.

5. A method according to claim 1, including further steps, prior to said measuring and comparing steps, of:

filtering the patient'"'"'s electrogram through a bandpass filter having a center frequency of about 25 Hz; and

sampling said filtered electrogram at a predetermined rate to obtain said metrics.

6. A method according to claim 5, wherein:

said filtering of the patient'"'"'s electrogram is performed through a filter having a pass band of 2 to 30 Hz.

7. A method according to claim 6, wherein said sampling step includes storing a peak value of said electrogram, and wherein said predetermined rate is about 80 Hz.

8. A method according to claim 6, wherein said storing step includes a sub-step of:

calculating and storing an NSR-mean value for each of the metrics of said first set, said NSR-mean value for each metric comprising the mean value of the metric during normal sinus rhythm of a normal electrogram.

9. A method according to claim 8, wherein said comparing step includes a sub-step of subtracting from each of the metrics in said second set of metrics the corresponding NSR-mean value therefor to provide NSR-mean-removed metrics corresponding to the patient'"'"'s heart rhythm.

10. A method according to claim 9, wherein said comparing step includes a further sub-step of classifying a heartbeat as either normal or abnormal based on the value of one of said NSR-mean-removed metrics exceeding a predetermined threshold value.

11. A method according to claim 10, wherein said diagnosing step includes a sub-step of diagnosing the patient'"'"'s heart rhythm as ventricular fibrillation if at least four heartbeats of ten successive heartbeats are classified as abnormal.

12. A method according to claim 1 wherein said diagnosing step includes a sub-step of diagnosing the patient'"'"'s heart rhythm as sinus tachycardia if at least a specified number of a predetermined greater number of last heartbeats are classified as normal.

13. A method according to claim 1, wherein said diagnosing step includes a sub-step of diagnosing the patient'"'"'s heart rhythm as sinus tachycardia if at least 8 of the last 10 heartbeats are classified as normal.