Method and apparatus for arrhythmia analysis of ECG recordings

• US 4,336,810 A
• Filed: 09/30/1980
• Issued: 06/29/1982
• Est. Priority Date: 09/30/1980
• Status: Expired due to Term

• Current Assignee: Del Mar Avionics
• Original Assignee: Del Mar Avionics
• Inventors: Ripley, John A., Cherry, Isaac R., Anderson, Donald L., Tanaka, David T.
• Primary Examiner(s): Kamm, William E.
• Application Number:
  US06/192,192
• Time in Patent Office: 637 Days
• Field of Search: 128/702, 128/703, 128/704, 128/710, 128/711, 128/712, 360/33 ME, 364/417
• US Class Current: 600/515
• CPC Class Codes: A61B 5/337 Playback at speeds other than the recording speed
• CPC Class Codes: A61B 5/349 Detecting specific parameters of the electrocardiograph cycle
• 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: A61B 5/7267 involving training the classification device
• CPC Class Codes: G11B 15/48 Starting; Accelerating; Decelerating; Arrangements preventing malfunction during drive change
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

Method and apparatus for arrhythmia analysis of Holter-type ECG recordings including a tape playback unit having an analog signal output of successive ECG complexes or heart beats, a converter for generating data composites representing each complex, a computer for making and storing data templates in bins corresponding to operator classification of complexes as normal, supraventricular ectopic (SVE), ventricular ectopic (VE), or unknown, for comparing successive composites with stored templates and conditions to determine a match condition, and for signaling a corresponding event to an event counter and display. In learn mode the computer signals the tape playback to stop and to display an unmatched complex for operator classification and possible subsequent generation of a new template, after which scanning is resumed.

Claims

1. In a method for the analysis of a tape recording of ECG signals for normal and arrhythmic events, the steps of scanning the tape recording at high speed to create an electrical signal comprising a succession of complexes representing successive heart beats, generating composites comprising a plurality of digital words which represent measured parameters characterizing each complex, comparing each successive composite with a plurality of known templates or conditions based on prior classifications of composites of the same tape recording and outputting an event signal if a match between a composite and a template or a condition is found, counting event signals corresponding to specific normal or arrhythmic events, and continuing to scan the tape until a no-match is found between the composite and any template or condition, stopping the scan and presenting a CRT analog display of the complex corresponding to said unmatched composite, classifying the complex, forming and storing an additional template based on the classification, incrementing the count of event signals according to the classification and continuing to scan the tape to completion.

2. The method as in claim 1 in which the step of generating digital words includes synthesizing a digital word for each selected parameter of the complex selected from among QRS width, R-sign, amplitude, pulse, positive area, negative area, and number of slope changes.

3. The method as in claim 2 in which the step of making templates from composites includes adding tolerances to the composite digital words.

4. The method as in claim 3 in which the digital words to which tolerances are added are those relating to positive area measurement, negative area measurement and QRS width.

5. The method as in claim 1 further in which the step of making the templates from composites includes adding tolerances to the composite digital words.

6. The method as in claim 1 in which the step of forming templates includes the step of creating a respective memory location bin for template classifications designated as normal, VE, SVE, and unknown.

7. The method as in claim 1 further including the step of rescanning the tape recording a second time and displaying a count of selected matched events for improved accuracy over said first scan.

8. A method as in claim 7 in which said selected events include total heart beats, VE, SVE, and unknown occurrences.

9. The methodas in claim 1 in which the step comparing with known templates includes classifying the templates as normal, VE, SVE, or unknown.

10. The method as in claim 1 further including the step of measuring instantaneous and average heart rate and using the same in combination with template classification to form a condition for subsequent composite classification.

11. The method as in claim 10 in which the template classifications include normal, SVE, VE, and unknown.

12. The method as in claim 11 further including the step of automatically classifying a composite SVE whenever a composite match is normal and an ectopic prematurity condition is simultaneously present based on the instantaneous and average heart rate, and subsequently incrementing SVE counts whenever such condition is present.

13. The method as in claim 11 further including the step of automatically classifying a composite VE whenever a composite template match is not normal and a prematurity condition is satisfied based on instantaneous heart rate and the number of immediately previous VE classifications, and incrementing the VE counts accordingly.

14. The method as in claim 1 in which the high-speed scanning step is carried out at from 60,120 or 240 times real time.

15. A method as in claim 1 further including the step of displaying successive complexes during said high-speed scanning step as a superimposed AVSEP display.

16. A method as in claim 1 in which the display of complexes corresponding to unmatched composites is a repeated time lapse display of an interval of a tape recording containing said unmatched complex.

17. A method as in claim 1 in which displaying of the unmatched complex includes the steps of storing a selected interval of ECG signal from the tape recording in a memory, and replaying that interval repeatedly on an oscilloscope.

18. The method of claim 17 in which the step of continuing the scanning of the tape includes the step of measuring the time occurrence of each unmatched complex on the tape recording and beginning said continuation of scan step at a preset time later in said recording and immediately following the said unmatched complex.

19. A method for the high-speed analysis of ECG data wherein said data is collected from a patient in actual time and stored for subsequent processing, and wherein said analysis includes recognition of at least VE beats and SVE beats, comprising the steps of:

measuring preselected parameters of a first plurality of normal beats from the stored patient ECG data and separately storing the same to form a first template bin composed of sets of preselected parameters of a series of normal beats,

measuring preselected parameters of a second plurality of VE beats from patient ECG data and separately storing the same to form a second template bin composed of sets of preselected parameters of a series of VE beats,

measuring preselected parameters of a third plurality of SVE beats from the stored patient ECG data and separately storing the same to form a third template bin composed of sets of preselected parameters of a series of SVE beats,

comparing the preselected parameters of each successive beat with the preselected parameters of each template in said first template bin to determine which of said beats is normal and not-normal,

comparing the normal beats with a prematurity condition to derive an (Auto) SVE event signal whenever the condition is satisfied and a normal event signal whenever the condition is not satisfied,

comparing the not-normal beats to pre-existing VE and to heart rate data conditions to determine which are auto VE beats, outputting a VE signal whenever said conditions are satisfied,

comparing the remaining not-normal beats with the preselected parameters of each of the series of SVE templates in said third template bin to determine which are SVE beats, and outputting an SVE event signal whenever a match is found,

comparing the remaining not-normal beats in said data with the selected parameters of the series of VE templates in said second template bin to determine which are VE beats, and outputting a VE signal whenever a match is found,

wherein said preselected parameters for each beat include at least QRS pulse width, plus area of the R wave, minus area of the S wave.

20. The method of claim 19 wherein said first, second and third template bins are characterized by upper and lower bounds for each of said preselected parameters as determined by said steps of measuring for said first, second and third plurality of beats respectively, said upper and lower bounds for each said parameters being the maximum and minimum values of said parameters corresponding to said first, second and third pluralities of beats corresponding to said first, second and third templates.

21. The method of claim 20 wherein said first, second and third pluralities of beats include said successive beats if said successive beats are characterized by values of said preselected parameters outside said upper and lower bounds pre-established by said first, second and third pluralities of beats;

and further comprising the step of;

re-establishing said upper and lower bounds for each parameter by including said value of said successive beats outside said pre-established upper and lower bounds.

22. In an apparatus for analyzing ECG tape recordings for arrhythmic events, means for scanning recordings at high speed and for creating an electrical signal of successive ECG complexes therefrom, means for receiving said signal and for converting each complex into a composite of digital words which represent measured parameters representing the same, means for making and storing a plurality of classified templates corresponding to previously-identified composites, means operative during said scan for comparing each new composite with said stored templates to determine if a match is found, means for stopping said scanning means and for displaying each complex corresponding to an unmatched composite in a stable view for operator classification, means responsive to operator classification for generating a new template for delivery to said storage means, means for counting and displaying a total of matched and classified events of said scanned tape recording.

23. Apparatus as in claim 22 in which said means for converting each complex is an analog computer and further in which said means for making and storing templates is a digital computer.

24. Apparatus as in claim 22 in which said means for converting each complex includes trigger means for sensing when a complex of predetermined strength and rise time is present, means responsive to said trigger means for measuring the QRS width of the complex from the inception of the initial R-wave to the nadir of the S-wave and for generating a digital representation thereof.

25. Apparatus as in claim 24 further including means responsive to the trigger means for measuring the magnitude of the time integral of the amplitude of the R-wave of the complex and for generating a digital representation thereof.

26. Apparatus as in claim 24 further including means responsive to said trigger means for measuring whether the initial R-wave of each complex is positive or negative and for generating a digital representation thereof.

27. Apparatus as in claim 24 further including means responsive to said trigger means for counting the number of slope changes of said video signal during said QRS width interval and for generating a digital representation thereof.

28. Apparatus as in claim 22 in which said means for storing templates includes a bin for each classification of complex including normal, SVE, VE and unknown.

29. Apparatus as in claim 28 in which said means for storing templates is a digital computer memory.

30. Apparatus as in claim 29 in which said means for comparing templates and composites is a digital computer including memory programmed to provide a means for each classification of complex and template including normal, SVE, VE, and unknown and further programmed to provide means for comparing each composite against stored templates in a decision matrix including:

normal, auto SVE, if normal and premature;

auto VE, if following a series of VE beats above a predetermined heart rate;

SVE;

VE, and unknown, and for providing an output signal for respective events, means for counting said classified events.

31. Apparatus as in claim 22, said means responsive to operator classification, includes a hand-held controller having a plurality of classification indices thereon responsive to operator touch to effect classification selection and to initiate restart of said scanning means.

32. Apparatus as in claim 22 further including means for computing the R--R interval between successive normal heart beats and for computing ectopic prematurity, average and instantaneous heart rate therefrom.

33. Apparatus as in claim 22 further including means responsive to operator control for generating a printout of the displayed complex.

34. An apparatus for analyzing extended tape recordings of a patient'"'"'s ECG activities for the presence of normal and ectopic beats, a tape playback transport including playback head means responsive to said tape recording for generating an analog electrical signal therefrom, means for supporting said tape recording for high-speed scan movement past said head so that successive complexes are detected, a display oscilloscope, means for generating AVSEP display on said oscillocope of said complexes during said high-speed scan of said tape recording, memory means for storing a selected interval of the signal on said tape recording for subsequent display in real time, means operative during said scan and responsive to said electrical signals for measuring preselected parameters of each of said complexes selected from among the QRS width or R-sign, slope, plus area, minus area, and amplitude, and for generating a digital composite of each complex therefrom, computer means for accepting each digital composite, and including memory means for forming template bins corresponding to patient data classified as normal, SVE, VE, and unknown events, such computer programmed to compare the data stored in template bins with each received composite and to generate an appropriate event whenever a match occurs, means responsive to a no-match condition for stopping the playback transporter causing the transfer of said oscilloscope to display at least a portion of the ECG interval from said memory means and including the unmatched complex therein, means for said computer further programmed to respond to operator input classification of said displayed complex to generate a new template for storage in a respective bin in said memory and for generating an event signal corresponding to such classification, counting means for totaling said matched and classified event signals.

35. A method for the high-speed analysis of ECG data wherein said data is collected from a patient in actual time and stored for subsequent processing, and wherein said analysis includes recognition of at least VE beats and SVE beats, comprising the steps of:

measuring preselected parameters of a first plurality of normal beats from the stored patient ECG data and separately storing the same to form a first template bin composed of upper and lower bounds of preselected parameters of a series of normal beats,

measuring preselected parameters of a second plurality of VE beats from patient ECG data and separately storing the same to form a second template bin composed of upper and lower bounds of preselected parameters of a series of VE beats,

measuring preselected parameters of a third plurality of SVE beats from the stored patient ECG data and separately storing the same to form a third template bin composed of upper and lower bounds of preselected parameters of a series of SVE beats,

comparing the preselected parameters of each successive beat with the preselected parameters of each template in said first template bin to determine which of said beats is normal and not-normal,

comparing the normal beats with a prematurity condition to derive an (Auto) SVE event signal whenever the condition is satisfied and a normal event signal whenever the condition is not satisfied,

comparing the not-normal beats to pre-existing VE and to heart rate data conditions to determine which are auto VE beats, outputting a VE signal whenever said conditions are satisfied,

comparing the remaining not-normal beats with the preselected parameters of each of the series of SVE templates in said third template bin to determine which are SVE beats, and outputting an SVE event signal whenever a match is found,

comparing the remaining not-normal beats in said data with the selected parameters of the series of VE templates in said second template bin to determine which are VE beats, and outputting a VE signal whenever a match is found, and

re-establishing each said upper and lower bound corresponding to said first, second and third pluralities of beats for each said parameter when a value of said parameter for a corresponding successive beat is outside said initially established upper and lower bounds, said successive beat corresponding to a normal beat, VE beat, or SVE beat.

36. In a method for the analysis of a recording of ECG signals for normal and arrhythmic events, the steps of:

scanning the recording at high speed to create an electrical signal comprising a succession of complexes representing successive heartbeats;

generating composites comprising a plurality of digital words which represent measured parameters characterizing each complex, each digital word having a tolerance associated therewith;

comparing each successive composite with a plurality of known templates based on said previously generated composites of the same tape recording; and

outputting an event signal if a match within said tolerances between a composite and a template is found,

whereby classification of said complexes is simplified by including all heartbeats as matched when said heartbeat falls within a predetermined tolerance of said templates.

37. The method as in claim 36 in which the digital words to which tolerances are added are those relating to positive area measurement, negative area measurement and QRS width.

38. In a method for the analysis in an automated apparatus of a recording of ECG signals for normal and arrhythmic events, the steps of:

scanning the recording at high speed to create an electrical signal comprising a succession of complexes representing successive heart beats;

generating composites comprising a plurality of digital words which represent measured parameters characterizing each complex;

measuring instantaneous and average heart rate and using the same in combination with template classification to form a condition for subsequent composite classification;

comparing each successive composite with a plurality of known templates or conditions based on said previously generated composites of the same recording;

automatically classifying a composite as auto VE whenever a composite template match is not normal and a prematurity condition is satisfied based on instantaneous heart rate and the number of immediately previous VE classifications; and

outputting an event signal if a match between a composite and a template or a condition if found whereby auto VE tachycardia is classified in an accelerated mode without stopping the analysis for nonautomated classification of each heartbeat.

39. An improvement in an apparatus for analyzing ECG recordings for arrhythmic events, comprising:

means for scanning recordings at high speed and for creating an electrical signal of successive ECG complexes therefrom;

means for receiving said signal and for converting each complex into a composite of digital words which represent measured parameters representing the same wherein said means for converting each complex is an analog computer for generating at least some of said measured parameters; and

means for making and storing a plurality of classified templates corresponding to previously-identified composites, wherein said means for making and storing templates is a digital computer.

40. The improvement of claim 39 wherein said analog computer integrates selected portions of said ECG complexes and wherein said digital computer compares said templates to said ECG complexes.