Method and system of ECG data review and analysis

• US 4,989,610 A
• Filed: 11/16/1987
• Issued: 02/05/1991
• Est. Priority Date: 11/16/1987
• Status: Expired due to Fees

• Current Assignee: Spacelabs, Inc. (OSI Systems Incorporated)
• Original Assignee: Spacelabs, Inc. (OSI Systems Incorporated)
• Inventors: Patton, Craig A., Sundahl, Roy E., Webert, Steven W., Dow, Harold G., Siegman, Craig S.
• Primary Examiner(s): Kamm, William E.
• Assistant Examiner(s): SHAY, DAVID M
• Application Number:
  US07/121,817
• Time in Patent Office: 1,177 Days
• Field of Search: 128/695-700, 128/702-705, 128/708-712, 364/413.01, 364/413.02, 364/413.03, 364/413.05, 364/413.06
• US Class Current: 600/508
• CPC Class Codes: A61B 5/337 Playback at speeds other than the recording speed
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/352 Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: G16H 15/00 ICT specially adapted for medical reports, e.g. generation or transmission thereof

Abstract

A method and computerized system for ECG data review and analysis are provided. In preferred embodiments, a transformation is performed on beat data and the transformed data are categorized into beat "bins." The clinician/user may select sensitivity to vary the number of bins created by the system. Once beats are preliminarily assigned to bins by the system, the clinician/user may review beats by bin and reassigned beats as desired. In preferred embodiments, recategorization is accomplished by "dragging" a beat to a new bin with the aid of a pointing device or "mouse." In preferred embodiments, a superimposition mode allows dynamic review of beat graphs with automatic detection, color coding, and optional pause at abnormal beats. Audio information is also preferably provided to aid the clinician/user in detecting abnormal beats. A P-wave marker capability is also provided to give a constant reference point during beat review in the superimposition mode. Preferred embodiments also provide the clinician/user immediate access to surrounding real-time data when reviewing individual beats.

Claims

1. A method of analyzing data representing a patient'"'"'s heart beats over a period of time as a plurality of waveforms, the method comprising the steps of:

(a) performing a Fourier analysis on the beat data to obtain pluralities of points in the complex plane representing the natural complex frequencies of each beat and preliminarily categorizing the beats into a plurality of bins of similar waveforms based on the proximity of corresponding pluralities of points in the complex plane;

(b) displaying a plurality of representative beats on a display screen, the beats being representative of a plurality of bins;

(c) selecting a bin for possible recategorization;

(d) after performing steps b and c, displaying the waveform of a representative beat from the selected bin superimposed over a waveform from a displayed bin to allow direct comparison of the waveform of the selected bin and the displayed bin;

(e) comparing the superimposed waveforms to determine if they are similar enough to be merged; and

(f) after comparing the superimposed waveforms, merging the selected bin and the displayed bin compared therewith if sufficient similarity is found.

2. The method of claim 1 including the additional step of analyzing the beats categorized within each bin to create a representative beat that is a composite of a plurality of beats within the bin.

3. A system for analyzing data representing a patient'"'"'s heart beats over a period of time as a plurality of waveforms, the system comprising:

means for performing a Fourier analysis on the beat data to obtain pluralities of points in the complex plane representing the natural complex frequencies of each beat and preliminarily categorizing the beats into a plurality of bins of similar waveforms based on the proximity of corresponding pluralities of points in the complex plane;

means for displaying the waveform of a representative beat which is representative of a plurality of bins;

means for selecting a bin for possible recategorization and displaying the waveform of a representative beat from the selected bin superimposed over a waveform from a displayed bin to allow direct comparison of the waveforms of the selected bin and the displayed bin;

pointing means for designating whether the selected bin and the displayed bin should be merged; and

means responsive to the pointing means for merging the selected bin and the displayed bin compared therewith.

4. The system of claim 3 wherein the waveforms for the representative beats of each bin comprise a composite waveform generated from the individual beats in the bin.

5. A method of analyzing data representing a patient'"'"'s heart beat over a period of time as a plurality of waveforms, the method comprising the steps of:

(a) performing a discrete Fourier analysis on the beat data to obtain pluralities of points representing the natural complex frequencies of each beat and preliminarily categorizing the beats into a plurality of bins of similar waveforms based on the proximity of corresponding pluralities of points in the complex plane;

(b) displaying the waveform of a beat which is representative of a plurality of bins on a display screen;

(c) selecting as a current bin one of the bins for which a representative beat is being displayed;

(d) selecting a beat waveform within the current bin as a current beat; and

(e) simultaneously displaying a representative waveform for the displayed bins and the current waveform and a plurality of waveforms surrounding the current waveform in a strip format to enable review of the heat activity at the time of the current waveform.

6. A method of analyzing data representing a patient'"'"'s heart beat over a period of time as a plurality of waveforms, the method comprising the steps of:

converting the waveforms to a plurality of points in the complex plane using a discrete Fourier transform;

comparing the corresponding Fourier transform-generated points in the complex plane to one another by determining whether one point in the complex plane falls within a window area of predetermined size which is centered about the other corresponding point in the complex plane; and

based on the results of the comparison, categorizing the beats corresponding to the coordinates into a plurality of bins of waveforms having similar characteristics.

7. A method for categorizing ECG data which contains a plurality of heartbeat waveforms, each waveform having a predetermined feature, the method comprising the steps of:

(a) sampling the data;

(b) detecting the occurrence of the predetermined feature in the sampled data and identifying each of the waveforms containing the predetermined feature;

(c) selecting each portion of the ECG data having a predetermined relationship to the predetermined feature of each of the waveforms;

(d) performing a discrete Fourier transform on each of the data portions to produce a plurality of points in the complex plane characterizing the natural complex frequencies of each data portion;

(e) categorizing each plurality of points in the complex plane into one of a plurality of bins by determining whether each point in the complex plane falls within a corresponding window area of predetermined size that is centered about a corresponding point that characterizes the natural complex frequencies of a waveform that represents the bin; and

(f) identifying each of the samples portions of ECG data with the bin into which the corresponding plurality of points in the complex plane is categorized.

8. The method of claim 7 wherein step (e) further comprises identifying all bins for which each of the plurality of points characterizing the natural complex frequencies of each data portion lies within each of the window areas located around the corresponding points that characterize the natural complex frequencies of a waveform that represents the bin and defining a new bin if no existing bin can be identified with the transformed data portion.

9. The method of claim 8 wherein a waveform portion is categorized in the bin identified with the transformed data portion which minimizes the sum of the magnitude of the errors between points defining the transformed data portion and the corresponding points that characterize the natural complex frequencies of a waveform that represents the bin.

10. The method of claim 7 the predetermined size of the window area is determined by a sensitivity criterion of acceptable similarity between two waveforms.

11. The method of claim 7 wherein, if any of the sampled data portions can be identified with more than one of the bins, step (e) further comprises categorizing the sampled data portion into the bin that minimizes the sum of the magnitudes of the differences between the points defining the corresponding transformed portion and the points that characterize the natural complex frequencies of a waveform that represents the bin.

12. The method of claim 11 wherein, if any of the sampled data portions is not identified with any of the existing bins, it is identified with a bin defined by the points that characterize the natural complex frequencies of the sampled data portion.

13. A method for categorizing ECG data which contains a plurality of heartbeat waveforms, each waveform having a predetermined feature, the method comprising the steps of:

(a) sampling the data with a uniform sampling frequency;

(b) detecting the occurrence of the predetermined feature in the sampled data and identifying each of the waveforms containing the predetermined feature;

(c) selecting each portion of each of the ECG data having a predetermined relationship with the predetermined feature of each of the waveforms;

(d) performing a discrete Fourier transformation of each of the data portions and producing a plurality of points in the complex plane corresponding to the natural complex frequencies of each of the transformed portions;

(e1) categorizing each plurality of points in the complex plane into one of a plurality of bins by comparing the location of each of the plurality of points in the complex plane with an area defined by a window located around the corresponding one of each of the plurality of points in the complex plane that characterizes the natural complex frequencies of a waveform that represents the bin;

(e2) identifying all bins for which each of the plurality of points defining the transformed portion lies within each of the areas defined by the windows located around the corresponding points that characterize the natural complex frequencies of a waveform that represents the bin, the sampled portion being categorized in the bin identified with the transformed portion which minimizes the sum of the magnitude of the errors between points defining the transformed waveform portion and the corresponding points that characterize the natural complex frequencies of a waveform that represents the bin;

(e3) categorizing the sampled waveform with a new bin if no existing bin can be identified with the transformed portion, the new bin being characterized by the natural complex frequencies of a waveform that represents the bin; and

(f) identifying each of the sampled waveforms with the bin with which the corresponding transformed portion is categorized.