US 5,655,540 AGrant
Noninvasive Method for Identifying Coronary Artery Disease Utilizing Electrocardiography Derived Data
Issue Date:1997-08-12
•30 Claims
•8 Drawing Sheets
Abstract
A method of analyzing empirically derived electrocardiograph (ECG) data which allows surprisingly accurate catagorization of subjects into various abnormal and normal classifications is disclosed. The presently preferred embodiment of the present invention applies an algorithm which compares root-mean-square (RMS) mean values derived from analysis of a representative composite of selected portions of a number of ECG PQRST waveforms obtained from (ECG) investigation of a subject, to similarly derived RMS mean and RMS standard deviation values present in a compiled data bank derived from (ECG) investigation of numerous normals, in each of a plurality of frequency range bands. A highly diagnostic numerical "Score" is calculated by addition of "Score" components found to be acceptable under certain mathematical criteria, and provided by the algorithm. Visually interpretable power spectral density plots supplement the method.
Metadata
Inventor
- Ronald D. Seegobin
Application Information
Application Number:US 4181751
Filing Date:1995-04-06
Priority Date:1995-04-06
Art Unit:335
Classifications
IPC:
A61B 504
Field of Search:
364128413.06702;705
Patent Drawings (8 sheets)
Description
Technical Field
The present invention relates to safe noninvasive systems and methods of use thereof for application in identifying coronary artery disease (CAD), which methods and systems are suitable for application in investigation of human subjects. More particularly the present invention is primarily a method of processing data derived from application of an electrocardiograph (ECG) system which involves mathematical and statistical techniques, data filtering and windowing, and application of a unique algorithm to the end that a highly predictive, easily interpreted numerically precise result is determined.
Background
It is generally accepted that approximately one-quarter of North Americans have some degree of Coronary Artery Disease (CAD). It is also generally accepted that approximately half thereof are not reliably detectable by conventionally applied diagnostic techniques, such as noting contour changes in the S-T segments of electrocardiograms (ECG's). The problematic nature the situation poses is perhaps most critically apparent when one considers that perioperative complications can be much more prevalent and serious in a patient with (CAD) than in a normal patient who does not have (CAD). That is, knowledge that a patient has (CAD) can be critical in fostering morbidity and mortality reduction procedure planning and scheduling on the part of medical professionals. As well, detection of (CAD) is, of course, important in everyday matters such as the planning and execution of a simple exercise routine.
It is noted that conventional (ECG) analysis provides time domain graphical results based primarily upon the low frequency (e.g. 0 to 40 Hz), content of a subject's cardiac signals monitored by an (ECG) system. This is the case whether a Frank orthogonal X-Y-Z; a standard 12 Lead or a multi-Lead monitoring etc. (ECG) system is utilized. For instance, a Patent to Brown et al., U.S. Pat. No. 5,077,667 describes a method for measuring a clinically useful characteristic of a fibrillating heart related to the elapsed time since the onset of ventricular fibrillation. Their significant variable is the power in the frequency range of 7 to 8 Hz. The Brown et al. Patent describes the use of a transformation of sampled analog time domain signals into the frequency domain and subsequent analysis thereof as a step intermediate to applying corrective treatment to a fibrillating heart.
A U.S. Pat. No. 4,680,708 to Ambos et al., describes the use of a Fast Fourier Transform (FFT) applied to a portion of an (ECG) cycle. Mathematical analysis of the last forty (40) milliseconds of a waveform derived from the time domain (QRS) complex allows for calculation of a Figure Of Merit, (FOM), based upon the frequency content thereof. Said (FOM) is correlated to the likelihood of a patient experiencing ventricular tachycardia. While the Ambos et al. Patent mentions the presence of high frequency components in a signal derived from the (ECG), said Patent primarily focuses upon the analysis of frequencies between 20 and 50 Hertz in arriving at the (FOM). The Ambos et al. Patent further states that "Recent studies . . . have used a variety of low (25 to 100 Hz) and high (250 to 300 Hz) band pass filters. A major limitation . . . is a lack of a-priori knowledge of the frequency distribution of signals of interest and the inherent risk that filtering will exclude signals of particular interest."
Other recent investigation has focused upon the diagnostic capability inherent in the presence of particular high frequency components present in an (ECG) signal. For instance, a very recent paper by Aversano et al., titled "High Frequency QRS Electrocardiography In The Detection Of Reperfusion Following Thrombolytic Therapy", (see Clinical Cardiology, (17, 175-182, April 1994)), states that the amplitude of the high frequency components, (e.g. 150-250 Hz), of the (QRS) complexes decreases during cardiac ischemia, and returns to normal with resolution thereof. It is also stated that high frequency electrocardiography is a rapid and reliable bedside technique for discriminating between successful and failed reperfusion in patients treated with thrombolytic agents for myocardial infarction. The Aversano et al. paper also states that "Studies involving high-frequency QRS electrocardiography are few and modest."
A paper by Moss and Benhorin titled "Prognosis and Management After a First Myocardial Infarction", New England J. Medicine, Vol. 322, No. 11, 1990 points out the importance of being able to identify and distinguish patients with various types of (CAD) so that appropriate treatment can be prescribed. This paper, in conclusion acknowledges that noninvasive techniques currently available for detecting jeopardized ischemic myocardium are imperfect.
The above sampling of relevant prior reference materials shows that techniques such as direct morphologic analysis of conventional time domain signals, application of (FFT) to (ECG) time domain derived signals to provide frequency domain spectra for analysis, analysis of high frequency components of (ECG) signals and the focusing on specific portions of a QRS complex etc. are known. There remains, however, need for additional and more probative noninvasive methods of analyzing (ECG) derived data which allow incipient (CAD) in patients to be identified with improved certainty. In particular, there is a need for a method of accurately identifying subjects with (CAD), the validity of which has been shown to provide utility by actual clinical testing.
The present invention provides an improved method of analyzing (ECG) derived data which has been shown by actual test, (in view of an extensive data bank accumulated by the inventor containing both normal and abnormal (ECG) data), to enable greatly improved ability to accurately and noninvasively separate abnormal from normal cardiac subjects. The method of the present invention, for instance, routinely allows identification of subjects with truely silent (CAD), and subjects who do not present with the tell-tale nonspecific S-T and T wave changes. The method of the present invention also routinely allows identification of subjects with nonspecific S-T and T wave changes, and allows identification and separate classification of subjects with prior myocardial infarction, abnormal patients who present with normal (ECG), and simultaneously distinguishes the population of abnormal subjects who present with normal (ECG).
Disclosure of the Invention
The present invention, in its presently preferred embodiment, utilizes a Frank orthogonal X-Y-Z Lead electocardiograph (ECG) system, but is primarily a method of analyzing and categorizing individual subject electrocardiogram (ECG) data. Said method has been shown to be capable of identifying and classifying subjects into cardiac categories of:
1. Normal, and
2. Abnormal:
a. Presents with prior myocardial infarction,
b. Presents with nonspecific S-T and T wave changes,
c. Presents with normal (ECG) but known otherwise to be abnormal.
As a starting point the present invention requires a substantial data base from which (ECG) data attributable to a "normal" population can be derived and used to form a "template" against which unknown subject (ECG's) can be compared. The present invention provides that by application of a discriminant Algorithm, (see supra), these unknown subjects may be appropriately classified. In the presently preferred embodiment of the present invention such a data base was developed by selection and testing of fit and healthy, relatively young subjects from families with a low prevalence of, and low risk factors for, coronary artery disease (CAD). Suitable subjects were required to fill out a questionnaire, analysis of which aided in determination of subject suitability as a "normal". A total of two-hundred-fifty (250) normal subjects were identified and (ECG) data obtained from each thereof. A random sampling of data from one-hundred-forty (146) subjects from said group of two-hundred-fifty (250) normals was assembled and analyzed to provide relevant, (see supra), composite root-mean-square, (RMS), mean and standard deviation values.
To arrive at said relevant normal RMS mean and RMS standard deviation values, (ECG) signals for each normal were derived by acquiring a number of, (typically one-hundred (100)), full (ECG) cycles, (i.e. full PQRST (ECG) cycles), for each normal subject, sampling each full cycle to provide six-hundred (600) data points over the extent thereof, and then selecting out the corresponding data points in each QRS complex in each of said full PQRST cardiac cycles. A single averaged (ECG) cardiac cycle was mathematically constructed from said number of QRS complexes and RMS mean and RMS standard deviation values calculated therefore.
In addition, similar RMS means and RMS standard deviation values were obtained from the same data, but which data had been subjected to digital filtering employing a Blackman-Harris window. The results are identified in the following table for each of the three Frank orthogonal (ECG) X-Y-Z system lead signals:
where SD stands for Standard Deviation.
Said resulting normal RMS mean and RMS standard deviation values for each of the frequency ranges and utilized Frank (ECG) system X-Y-Z lead serve to define assumed Gaussian "templates" for normals, against which similarly derived RMS means, acquired from individual subjects, are compared under the guidelines of the Algorithmic Method of the present invention, (see supra).
To date data from more than one-thousand (1000) abnormal subjects has been assembled by the inventor, and subject RMS mean values derived therefrom. It is noted that abnormal subjects tested provide representatives from each the three abnormal groups identified infra.
Continuing, to apply the Algorithm Method of the present invention many, (typically one-hundred (100)), of human subject derived full PQRST (ECG) cardiac cycles are obtained. Said full (ECG) cardiac cycles are each then sampled to provide six-hundred data points over the extent thereof and the sampled data points corresponding to the QRS complex in each full cycle are mathematically averaged to provide a composite QRS complex. As well, the digital filtering and application of the Blackman-Harris Window to allow calculation of the RMS means which correspond to each frequency range and Frank (ECG) system X-Y-Z lead utilized, are performed. The end result can be expressed as a table of data, (not shown), similar to the table presented above for the normal data but which contains only composite RMS mean values.
The Algorithm employed in the method of the present invention embodiment then provides for a maximum of thirty (30) calculations to be performed as follows:
a. Up to Fifteen of said calculations involve calculating the difference between the subject composite RMS mean and the corresponding normal composite RMS mean, and dividing the result by the corresponding normal RMS standard deviation, for each Frank X-Y-Z (ECG) lead and each frequency range band, (i.e. the fifteen calculations break down as five (5) frequency range bands per each of the Frank X-Y-Z (ECG) leads.)
b. Twelve of said calculations involve finding the RMS ratio for each frequency range band, (e.g. 0-10 Hz, 10-60 Hz, 60-150 Hz and 150-250 Hz) to the total sum of all said frequency band RMS contributions for each appropriate said Frank (ECG) X-Y-Z lead, for a subject, and subtracting therefrom equivalent RMS ratio mean results derived based upon the same calculations as applied to normal data, and dividing the result by the RMS standard deviation for the corresponding frequency range, of said normal data.
c. Three of said calculations involve calculating the difference between ratios of the subject composite RMS means of Frank lead X/Y, Y/Z and X/Z ratios and the corresponding RMS means of normal X/Y, Y/Z and X/Z ratios and dividing by the RMS value of the normal RMS standard deviations of said ratios.
Each of the above identified thirty (30) calculations will result in a number (Pi). A "Score" component number (Si) is then derived based upon where a (Pi) number lies in an assumed Gaussian Distribution. This is calculated based upon normal data RMS Means and RMS Standard Deviations (X) as follows:
Each of the resulting subject RMS mean values associated with a calculated Si value is then analyzed to determine if it is less than or equal to ninety-five (95%) percent of the data points from which the normal RMS mean was calculated. If this is the case the associated Si is accepted. Otherwise it is rejected. That is, a ninety-five (95%) confidence interval, based upon normal data spread, is imposed, in determining whether to accept a calculated (SI) value.
Accepted Si values are then selected and added together to provide a final numerical "Score". (Note, two Scores in the catagory a. above are commonly not selected as being redundant to other Scores. Said commonly unselected Scores are better identified in the Detailed Description Section).
It has been found that if said final numerical "Score" is "low" (e.g. approximately 0 to 7) then the subject is more likely to be normal. If the final numerical "Score" is high (e.g. greater than about 8), then the subject is more likely to be abnormal. For instance, a "Score" of 7 provides a ninety (90%) percent confidence of normality, and a "Score" of 8.4 provides a ninety-five (95%) confidence of normality, (See FIG. 9).
As will be better presented in the Detailed Description Section of this Disclosure, the results of the application of the method of the present invention as described above can be presented in numerous ways. A particularly relevant approach is to present the results on a graph of ("Sensitivity" vs. "100-Specificity"). (The present invention provides that the "Score" value be plotted against the abscissa (100-Specificity) and that percentage of a group having said "Score" be plotted on the (Sensitivity) ordinate). Said approach to presentation is generally known as an ROC curve, (ROC stands for Receiver Operation Characteristic as the technique was originally derived for use in testing radio receiver quality). Said approach to presentation serves to visually demonstrate the success of the present invention method of analyzing (ECG) derived data. In particular, abnormal subjects which can not be identified by conventional (ECG) analysis techniques, are seen to be easily identified by application of the present invention algorithm.
In addition, the present invention provides that time domain data obtained from Frank X-Y-Z (ECG) leads should be subjected to a Fourier Transform and manipulated to provide Power Spectral Density (PDS) vs. Frequency plots. As will also be better presented in the Detailed Description Section of this Disclosure, said (PDS) plots are typically easier than associated (ECG) data vs. time plots to visually interpret. Said plots complement the above described "Scoring" system approach to identifying (CAD).
Summary
It is generally accepted that approximately one-quarter of North Americans have Coronary Artery Disease (CAD), and that approximately half thereof are not reliably detectable by conventionally applied diagnostic techniques.
Numerous researchers and inventors have investigated techniques for analysis of electrocardiogram (ECG) derived data in an attempt to overcome said identified deficiency. For instance, the various techniques of transformation of time domain signals to the frequency domain, focusing on portions of a full (ECG) PQRST cardiac cycle, extraction of high frequency content, (e.g. in the 150 to 300 Hz range), in (ECG) data for analysis etc. are known. However, no known approach has been wholly successful in reliably identifying, for instance, subjects with silent (CAD) who present with an apparently normal (ECG) waveform. Various reasons have been advanced as to why this remains the case, including the lamented lack of a-priori knowledge of the frequency distribution of (ECG) signals, and the existence of but only a few studies which focus on the high frequency content of the (ECG) QRS waveform.
The present invention has significantly overcome the above identified deficiencies. As a starting point for application of the present invention a rather extensive validated data base of (ECG) data with requisite fidelity, for both normal and abnormal subjects was accumulated. The presently preferred embodiment of the present invention obtained (ECG) data using a Frank orthogonal X-Y-Z (ECG) lead system, hence allowed simultaneous recording of three (mutually independent ECG) signals during application. Composite (ECG) signals in which the requisite frequency content was present, and in which filtering provided specific bands of frequency content were analyzed. Normal population Root-Mean-Square (RMS) signals for mean and standard deviation were calculated for (ECG) data containing the requisite frequencies, and for filtered bands, (e.g. 0-10 Hz, 10-60Hz, 60-150 Hz, and 150-250 Hz). High quality data was derived by obtaining a large number, (typically one-hundred), of full (ECG) PQRST waveforms from each normal subject tested, selecting out the portions thereof which correspond to the depolarization process, (the QRS complex), by filtering, (typically digital), and windowing techniques, and forming an average representative QRS waveform therefrom. In combination, filtering techniques to provide selected frequency bands as alluded to infra, and RMS signal processing techniques, then provided a composite RMS mean and RMS standard deviations data base for various frequency content ranges identified above, for a normal population. The present invention makes use of said normal population RMS mean and RMS standard deviation data as a "template", against which data from subjects are compared during use.
Practice of the present invention requires that (ECG) data be obtained from a subject in a manner essentially identical to the manner in which data are obtained from normals, as described above. Subject RMS mean values for full, and for limited band filtered frequency ranges, are then calculated. Again, as with the normal population this is accomplished by forming a representative average QRS waveform from many, (typically one-hundred (100)) subject QRS waveforms and utilizing filtering, windowing and RMS techniques to arrive at values for said RMS means. Subject RMS mean data are then compared to normal RMS mean and RMS standard deviation data by application of an algorithm. Said algorithm involves finding differences of subject and normal RMS mean values, presented as a multiple of corresponding normal RMS standard deviation values for each Frank X-Y-Z (ECG) system lead; as well as differences between subject and normal RMS mean values of ratios of data obtained from various (ECG) system leads as a multiple of normal data RMS Standard Deviations of said ratios for each Frank X-Y-Z (ECG) system lead, and involving similar ratio calculations based upon the relative RMS mean value associated with each frequency band for each lead observed. Resulting calculated numbers are then subjected to a truncating process by comparison to an assumed gaussian profile derived from normal data RMS Mean and RMS standard deviations, and a ninety-five (95%) percent confidence factor applied thereto to arrive at a set of accepted "Score" components which are added together to provide a diagnostic "Score". Said algorithm is better described in detail elsewhere in this Disclosure so will not be further discussed presently.
Continuing, it is to be understood and emphasized that the present invention is not to be considered to be limited in application to (ECG) systems which utilize Frank X-Y-Z (ECG) lead systems in (ECG) data acquisition, but rather is generally applicable to systems which utilize any number and type of leads. The presently preferred embodiment of the present invention is then not to be considered as limiting of the algorithmic approach to data analysis of the present invention, but is presented only as a presently preferred, and tested, example of the application thereof.
The method of the present invention is then found in an algorithm which compares (ECG) data obtained from (ECG) systems utilizing any number of leads, which algorithmic approach involves comparison of subject derived (ECG) data parameters to similar (ECG) data parameters which are representative of a normal population. The present invention method algorithm provides that (ECG) derived data should be subjected to filtering and windowing techniques to allow comparison of appropriate subject derived data parameters to corresponding normal (ECG) derived data parameters in any of a plurality of frequency ranges, and that ratios of so derived appropriate subject and normal (ECG) derived data parameters should be included in arriving at a numerical "Score" with which is associated a highly reliable diagnostic capability.
In addition, the present invention utilizes frequency domain transformed power spectral density plots as visual aids to compliment interpretation of the "Scoring" system results.
It is therefore a purpose of the present invention to teach that a preferred approach to analysis of empirically derived subject (ECG) data begins with the accumulation of a significant data base which documents various parameters associated with a normal population of subjects.
It is another purpose of the present invention to teach that a preferred approach to analysis of empirically derived subject (ECG) data involves simultaneous comparison of similar various subject and normal parameters derived from analysis of (ECG) data under the guide of a specific algorithm.
It is still yet another purpose of the present invention to teach that filtering and windowing techniques should be applied to empirically derived normal and subject (ECG) data to allow independent consideration of comparisons of subject and normal parameters derived from analysis of (ECG) data based upon information present in specific frequency band ranges.
It is yet still another purpose of the present invention to teach that appropriate subject and normal parameters for comparison in a preferred approach to analysis of empirically derived subject (ECG) derived data are the RMS mean and RMS standard deviation values calculated from empirically derived subject and normal data in various frequency band ranges.
It is another purpose of the present invention to teach that specific portions of a full (ECG) PQRST waveform should be selected for use in deriving appropriate subject and normal parameter values for comparison by application of a specific algorithm and that the QRS depolarization complex is best suited for analysis when high (e.g. above approximately 100 Hz), frequency content is of interest.
It is still another purpose of the present invention to teach that an appropriate algorithm for analysis of empirically derived subject (ECG) data should provide for simultaneous comparison of specific subject parameters to similarly derived normal parameters, and ratios thereof.
It is yet another purpose of the present invention to teach that frequency domain transformed power spectral density plots can be used as visual aids to interpretation of the "Scoring" system results, as well as plots of (ECG) signal amplitude in the time domain.
Brief Description of the Drawings
FIG. 1a shows a representation of a human torso with Frank X and Y (ECG) system leads attached thereto.
FIG. 1b shows a cross section taken at a--a in FIG. 1, with Frank Z (ECG) system leads attached thereto.
FIG. 2 shows a demonstrative "PQRST" (ECG) waveform.
FIG. 3 demonstrates a table for recording data necessary for practice of the present invention.
FIG. 4 shows an assumed gaussian distribution for use in assigning "Score" component numbers during practice of the present algorithm.
FIG. 5 shows sample results as provided by practice of the present invention plotted on an ROC curve in which a (100-Specificity), and Sensitivity, appear on the abscissa and ordinate respectively.
FIGS. 6X1 through 6Z4 show twelve related plots of subject time domain sample date recorded from Frank X, Y and Z leads. Higher frequency band data are presented as one progresses from FIGS. 6X1 to 6X4, and from FIGS. 6Y1 to 6Y4, and from FIGS. 6Z1 to 6Z4.
FIGS. 7aX1 through 7aZ5 show fifteen related plots of subject frequency domain power spectral density. Shown in FIGS. 7aX1, 7aY1 and 7aZ1 are transforms of full frequency band requisite data. Transforms of data from progressively higher frequency bands are presented as one progresses from FIGS. 7aX1 to 7aX5, and from FIGS. 7aY1 to 7aY5 and from 7aZ1 to 7aZ5.
FIGS. 7bX1 through 7bZ5 show fifteen related plots of subject time domain sample data. Shown in FIGS. 7bX1, 7bY1 and 7bZ1 are full frequency band requisite data. Data from progressively higher frequency bands are presented as one progresses from FIGS. 7bX1 to 7bX5 and from 7bY1 to 7bY5 and from 7bZ1 to 7bZ5.
FIG. 8 shows actual data obtained from various subject groups by practice of the present invention, on a ROC curve
FIG. 9 shows actual data obtained from various subject groups by practice of the present invention, on a graph in which the abcsissa is scaled linearly with the present invention "Score".
FIG. 10 shows a flow chart of the steps of the preferred embodiment of the method of the present invention.
Detailed Description
In the following a specific embodiment of the present invention is presented. Said specific embodiment assumes the use of an (ECG) system which utilizes Frank (ECG) orthogonal X-Y-Z leads. It is to be understood, however, that the present invention is not limited to such and can be practiced with (ECG) systems in which any number of leads, (e.g. standard twelve (12), sixteen (16), or mapping arrays of twenty-four (24) or more etc.), are present, and in which only some of the present leads are utilized. The following specific embodiment is presented as it is well documented and is presently the preferred embodiment.
Turning now to the drawings, there is shown in FIG. 1(a) a frontal view of a torso of a human, with (ECG) Frank X and Y leads properly affixed thereto. FIG. 1(b) shows a cross section taken at a--a in FIG. 1(a) with (ECG) Frank Z leads properly attached thereto. In use said (ECG) Frank X-Y-Z leads are attached to an (ECG) system and serve to effect orthogonal monitoring of (ECG) full cardiac cycle PQRST signals which are essentially shaped as shown in FIG. 2.
The present invention requires as a starting point that a significant data base be available, which significant data base contains representative composite (ECG) data for all, or some portion of full (ECG) PQRST cycles for each (ECG) lead, for a normal population. (Note, a normal population is defined as one in which the subjects have no detectable coronary artery disease (CAD) by history and multiple conventional diagnostic tests, and are not at risk therefore based upon age, family history etc.) Such a significant data base for normals was acquired by obtaining a number of full (ECG) PQRST cycles from each of the Frank X-Y-Z (ECG) leads present in the presently discussed embodiment of the invention, from each of two-hundred-fifty (250) normals. (It is noted that the present invention is not limited to cases in which all leads present in an (ECG) system are monitored but that a preferred embodiment does utilize all available information). Next, a random sample of one-hundred-forty-six (146) of said two-hundred-fifty (250) normals was selected and a representative number of the full (ECG) PQRST cycles from each, (typically one-hundred (100)), for each Frank X-Y-Z (ECG) lead, were then selected and each subjected to a sampling procedure which provided some number of data points for each, (six-hundred (600) was chosen in the presently discussed embodiment). Next, the sampled data points corresponding to the QRS depolarization complexes in each selected full (ECG) PQRST cycle were selected and a representative composite QRS complex for each Frank (ECG) X-Y and Z leads formed therefrom by mathematical averaging thereof. Said representative composite was then subjected to filtering and windowing techniques to provide a number of data sets for each of the Frank (ECG) X-Y-Z leads. Said data sets in the presently preferred embodiment of the present invention provide information present in said representative composite in the frequency bands:
a. All frequencies;
b. Between zero (0) and ten (10) Hz;
c. Between ten (10) and sixty (60) Hz;
d. Between sixty (60) and one-hundred-fifty (150) Hz;
e. Between one-hundred-fifty (150) and two-hundred-fifty (250) Hz.
For each of the Frank (ECG) X-Y-Z leads then, five (5) sets of data were derived as described, and from each of said sets of data a Root-Mean-Square (RMS) mean and (RMS) standard deviation were calculated. This, it will be appreciated, resulted in fifteen (15) RMS representative composite means and standard deviations being available, (five for each Frank (ECG) X-Y-Z lead).
In view of the described RMS mean and RMS standard deviations (SD's) available clinical application of the present invention can be practiced.
To practice the present invention, data are obtained from a subject in a manner essentially the same as described infra for normals. That is, a number of full (ECG) PQRST full cardiac cycles from each Frank (ECG) X-Y-Z lead are obtained and a representative number thereof are selected and subjected to a sampling procedure. Some portion of each full PQRST waveform is selected, (e.g. the QRS depolarization complex is utilized in the presently preferred embodiment of the present invention), and a representative composite thereof formed therefrom for each Frank (ECG) system X-Y-Z lead. For each representative composite a RMS mean is then calculated so that a table equivalent to that shown in FIG. 3, but containing subject RMS mean data, is formed.
With the described normal RMS mean and RMS standard deviation data, and subject RMS mean data then available, the algorithm of the method of the present invention can be applied to arrive at a diagnostic mathematical "Score".
The algorithm of the present invention involves mathematical comparison of:
a. Normal and subject RMS means in view of normal RMS standard deviation;
b. Ratios of normal and subject RMS frequency range band means to the summation of RMS means for all frequency range bands for each Frank (ECG) X-Y-Z lead in view of normal standard deviation for the numerator frequency band.
c. Ratios of normal and subject Frank (ECG) X-Y-Z lead RMS means in view of normal standard deviations of said ratios.
Briefly, application of each of the identified steps provides a numerical result (Pi), which in general is typically not a whole integer. The next step is to process said numerical result (Pi) by comparison to an assumed Gaussian Distribution derived from the normal population data to arrive at a whole number integer which represents how many RMS standard deviations the subject RMS mean is away from the normal RMS mean, and assign a whole integer "Score" component number (Si) based thereupon. The algorithm then requires that a ninety-five (95%) confidence interval, based upon normal RMS standard deviation data be applied to determine if a "Score" component should be accepted and included in calculation of a final "Score", said final "Score" being arrived at by an addition of accepted "Score" components. Said algorithm will now be described in detail.
The first step in applying the algorithm of the presently described presently preferred embodiment of the present invention is perform up to fifteen (15) calculations comprising subtracting the Subject RMS mean from a corresponding Normal RMS mean and dividing the result by a corresponding normal RMS standard deviation for each Frank (ECG) X-Y-Z lead in each frequency range identified infra, to provide numbers (Pi).
(Note, the accompany computer printout page labeled "avgasc" provides the various RMS Means and Standard Deviations referred to in the following, corresponding to the P1-P30 number.)
For the Frank (ECG) X lead, (I.E. HORIZONTAL AXIS):
For all frequencies: ##EQU1##
For the frequency range band zero (0) to ten (10 Hz): ##EQU2##
For the frequency range band ten (10) to sixty (60 Hz): ##EQU3##
For the frequency range band sixty (60) to one-hundred-fifty (150 Hz): ##EQU4##
For the frequency range band one-hundred-fifty (150 Hz) to two-hundred-fifty (250 Hz): ##EQU5## For the Frank (ECG) Y lead, (I.E. VERTICAL AXIS):
For all frequencies: ##EQU6##
For the frequency range band zero (0) to ten (10 Hz): ##EQU7##
For the frequency range band ten (10) to sixty (60 Hz): ##EQU8##
For the frequency range band sixty (60) to one-hundred-fifty (150 Hz): ##EQU9##
For the frequency range band one-hundred-fifty (150 Hz) to two-hundred-fifty (250 Hz): ##EQU10## For the Frank (ECG) Z lead, (I.E. FRONT TO BACK AXIS):
For all frequencies: ##EQU11##
For the frequency range band zero (0) to ten (10 Hz): ##EQU12##
For the frequency range band ten (10) to sixty (60 Hz): ##EQU13##
For the frequency range band sixty (60) to one-hundred-fifty (150 Hz): ##EQU14##
For the frequency range band one-hundred-fifty (150 Hz) to two-hundred-fifty (250 Hz): ##EQU15## *Note: In practice of the present invention the P5 and P13 data are at times not utilized as it has been found to be redundant.
Twelve (12) additional groups of calculations are then performed in which the relative RMS mean content of each frequency range band identified infra is determined as a percentage of the RMS means of the sum of the filter derived frequency range bands for each Frank (ECG) system X-Y-Z system lead, for both Subject and Normal data, the differences therebetween being divided by the corresponding normal RMS Standard Deviation to provide additional numbers (Pi):
For Frank (ECG) X lead, (I.E. HORIZONTAL AXIS:
Define: ##EQU16##
Then: ##EQU17##
Define: ##EQU18##
Then: ##EQU19##
Define: ##EQU20##
Then: ##EQU21##
Define: ##EQU22##
Then: ##EQU23##
Define: ##EQU24##
Then: ##EQU25##
Define: ##EQU26##
Then: ##EQU27##
Define: ##EQU28##
Then: ##EQU29##
Define: ##EQU30##
Then: ##EQU31## For Frank (ECG) Z lead, (I.E. FRONT TO BACK AXIS):
Define: ##EQU32##
Then: ##EQU33##
Define: ##EQU34##
Then: ##EQU35##
Define: ##EQU36##
Then: ##EQU37##
Define: ##EQU38##
Then: ##EQU39##
Three (3) additional calculations are then performed in which Subject RMS means of ratios of RMS means obtained from the Frank (ECG) X-Y-Z leads, are subtracted from corresponding RMS means of ratios obtained similarly from normals, the results of which subtraction are then divided by RMS Standard Deviations of normal corresponding RMS ratios to provide additional numbers (Pi): ##EQU40##
Continuing, each of the above thirty (30) calculated numbers:
(PX1-P1, PX2-P2, PX3-P3, PX4-P4, PX5-P5, PX6-P6 PX7-P7, PX8-P8, PX9-P9),
(PY1-P10, PY2-P11, PY3-P12, PY4-P14, PY6-P15, PY7-P16, PY8-P17, PY9-P18),
(PZ1-P19, PZ2-P20, PZ3-P21, PZ4-P22, PZ5-P23, PZ6-P24, PZ7-P25, PZ8-P26, PZ9-P27),
(P(X/Y-P28), P(Y/Z)-P29 and P(X/Z)-P30),
(generally identified as (Pi)),
is then compared to a corresponding assumed Gaussian distribution of normal data to arrive at a "Score" component number. If a number (Pi) is within some .+-."X" RMS Standard Deviation range of the RMS mean as shown below, a "Score" component number (Si) is taken to be:
FIG. 4 demonstrates this graphically.
Continuing, each of the resulting "Score" component numbers (Si) calculated as just described is then subjected to a final test to determine if it should be accepted or rejected. Said final test involves comparing the Subject RMS mean to the data from which the Normal RMS mean was calculated. If less than or equal to ninety-five (95%) percent of the data points from which the Normal RMS mean was calculated are more than the subject's RMS mean the associated "Score" component (Si) is accepted, otherwise it is rejected. Accepted "Score" component numbers are then added to provide a final numerical "Score".
It has been found that if said final numerical "Score" is "low", (e.g. approximately 0 to 7), then the Subject involved is more likely to be normal. If the final numerical "Score" is "high", (e.g. greater than about 8), then the Subject is more likely to be abnormal.
A particularly relevant approach to presenting the results of applying the disclosed method of the present invention is demonstrated by FIG. 5. FIG. 5 shows a plot in which the abscissa is (100-specificity) and the ordinate is (sensitivity). These terms are well known and mean: ##EQU41##
The curve in FIG. 5 is demonstrative of those which populations of subjects provide in an (ROC) format. The present invention method provides that (ROC) curves be prepared by associating a "Score" value with the abscissa, in a nonlinear manner, and the percentage of a group having said "Score" value with the ordinate of such a plot. The success of the present invention in identifying and distinguishing abnormal subjects has been demonstrated to be quite striking. FIGS. 8 and 9, discussed supra, better serve to demonstrate this with actual empirically derived data.
FIGS. 6X1-6Z4 show twelve (12) diagrams, 6X1, 6X2, 6X3, 6X4, 6Y1, 6Y2, 6Y3, 6Y4, 6Z1, 6Z2, 6Z3 and 6Z4, of typical subject time domain data as provided by practice of the present invention. Data for each Frank orthogonal X-Y-Z lead, (horizontally), and for the various filtered and windowed frequency range bands, (vertically--frequency band increases as progress down the page) are shown. (Note that in the higher frequency filtered plots a peak envelope is shown). The designation of "SAMPLE NUMBER" on the abcissa refers to a digital filter data point within a QRS complex. The abcissa is thus a time axis spanning the QRS complex which can be coidentified as a Sample Number taken at some point after the QRS complex begins.
More particularly, FIGS. 6X1, 6Y1 and 6Z1 show filtered composite subject (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system in frequency band range of 0.0-10 HZ. FIGS. 6X2, 6Y2 and 6Z2 show filtered composite subject (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system for the frequency band of 10-60 HZ. FIGS. 6X3, 6Y3 and 6Z3 show filtered composite subject (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system for the frequency band of 60-150 HZ. FIGS. 6X4, 6Y4 and 6Z4 show filtered composite subject (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system for the frequency band of 150-250 HZ. It is noted that the plots for the 60-150 and 150-250 HZ bands present as "envelopes" as signals go positive to negative and vice versa in very short time periods, (i.e. over a few "Sample Numbers"). All plots in FIGS. 6X1-6Z4 have the ordinate marked in micro-volts, and the abscissa is marked in digital filter data points 0 to 600, taken at progressive times during an (ECG) cycle.
FIGS. 7aX1-7aZ5 show fifteen (15) diagrams of typical subject data in frequency domain Power Spectral Density form, (with Magnitude on ordinate), plotted as a function of Frequency, (on abscissa). FIGS. 7bX1-7bZ5 show fifteen (15) diagrams of typical subject data in Time Domain form, (Magnitude on ordinate), plotted as a function of time, (on abscissa). All said identified plots provide Magnitude, on the ordinate, in microvolts.
More particularly, it is noted that FIGS. 7aX1, 7aY1 and 7aZ1 show subject composite (ECG) data set frequency domain power spectral density plots derived from X, Y and Z leads, respectively, of a Frank ECG system, over a frequency band of 0.0 to 100 HZ. FIGS. 7aX2, 7aY2 and 7aZ2 show subject composite (ECG) data set frequency domain power spectral density plots derived from X, Y and Z leads, respectively, of a Frank ECG system, over a frequency band of 0.0 to 15 HZ. FIGS. 7aX3, 7aY3 and 7aZ3 show subject composite (ECG) data set frequency domain power spectral density plots derived from X, Y and Z leads, respectively, of a Frank ECG system, over a frequency band of 0.0 to 80 HZ. FIGS. 7aX4, 7aY4 and 7aZ4 show subject composite (ECG) data set frequency domain power spectral density plots derived from X, Y and Z leads, respectively, of a Frank ECG system, over a frequency band of 50 to 200 HZ. FIGS. 7aX5, 7aY5 and 7aZ5 show subject composite (ECG) data set frequency domain power spectral density plots derived from X, Y and Z leads, respectively, of a Frank ECG system, over a frequency band of 100 to 300 HZ. All plots in FIGS. 7aX1-7aZ4 have the ordinate marked in micro-volts, and the abscissa is marked in HZ, (i.e. cycles per second).
As well, FIGS. 7bX1, 7bY1 and 7bZ1 show subject composite (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system in an unfiltered full requisite frequency band. FIGS. 7bX2, 7bY2 and 7bZ2 show subject composite (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system in a filtered frequency band range of 0.0-10 HZ. FIGS. 7bX3, 7bY3 and 7bZ3 show subject composite (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system in a filtered frequency band range of 10-60 HZ. FIGS. 7bX4, 7bY4 and 7bZ4 show subject composite (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system in a filtered frequency band range of 60-150 HZ. FIGS. 7bX5, 7bY5 and 7bZ5 show subject composite (ECG) data set time domain waveforms obtained from X, Y and Z leads, respectively, of a Frank ECG system in a filtered frequency band range of 150-250 HZ. It is noted that the plots for the 60-150 and 150-250 HZ bands present as "envelopes" as signals go positive to negative and vice versa in very short time periods, (i.e. over a few "Sample Numbers"). All plots in FIGS. 7bX1-7bZ5 have the ordinate marked in micro-volts, and the abscissa is marked in digital filter Sample Number data points, taken at progressive times during a (ECG) cycle. The present invention then makes use of such visual aids as an added feature. The three curves in each plot represent normal mean and plus/minus one standard deviations, and subject data. It is also to be understood that the above described approach to diagnosis can be applied to tracking patients over time and can be applied before and after various stress tests which attempt to provoke otherwise indolent or silent coronary artery abnormalities. Stress tests can, for example, involve treadmill exertion or a cold pressor test in which a subject simply places an arm into cold water for a few minutes. Changes in "Score" results combined with changes in the appearance of Power Spectral Density (PSD) and Amplitude Plots over time or before and after stress tests can provide insight as to a subject's coronary health not made available by less vigerous testing. Multiple mean curves can be simultaneously presented on a single plot to allow easy visual comparison of changes in Power Spectral Density as a function of time or stress. Observation of changes in (PSD) plots in the various frequency bands is a correlated part of the method of the present invention. Of particular interest, the inventor has noted that plots of (PSD) in the frequency ranges of sixty (60) to one-hundred-fifty (150) HZ and one-hundred-fifty (150) HZ to two-hundred-fifty (250) HZ show the greatest change in visually observable shape when a cold pressor test is administered. This is considered a significant observation.
Note also that as shown in FIG. 7aX1, for instance, it is common to include numerical representation in frequency as well as the time domain plots. Four numbers can be present. Using the Power Spectral Density plot as an example, when present said numbers are representations of:
Upper left--the number of Standard Deviations a Subject Power Spectral Density Value is away from a corresponding Normal Power Spectral Density Value for the Frequency Band in the Plot.
Lower Left--the Percentage of Normals which are below the Subject Power Spectral Density Value for the Frequency Band in the Plot.
Upper Right--the number of Normalized, (i.e. Subject Power Spectral Density Value in the Frequency Band of the Plot divided by the Sum of Power Spectral Density Values for all Frequency Bands), Standard Deviations a Subject Power Spectral Value is away from a corresponding Normalized Subject Power Spectral Density Value for Normals for the Frequency Band in the Plot.
Lower Right--the Percentage of Normals which are below the Normalized Subject Power Spectral Density Value for the Frequency band in the Plot.
Said numbers and visual Plots aid in interpretation of a Subject's data, but are of secondary importance to the "Score", calculated as described infra herein.
Next FIGS. 8 and 9 show (ROC) plots for actual data arrived at using the present invention method. Again, (ROC) curves typically plot Sensitivity vs. (100-Specificity) on ordinate and abscissa respectively, presented as percentages. Said Plots in FIGS. 8 and 9 were generated by associating the present invention "Score" with the abscissa (100-Specificity), but with the zero (0) thereof being at the right side so that the "Score" increases to the left. As the "Score" increases the percentage of each group of subjects associated therewith is plotted on the ordinate. By observation of FIGS. 8 and 9 it will be appreciated that as the "Score" increases the percentage of normals in a group of known normals having said "Score" value drops off rapidly, but the percentage of known abnormals in a group of known abnormals drops off much more slowly. For instance, at a "Score" of zero (0) all members of all groups are present. At a "Score" of five (5) approximately eighty (80%) percent of all members of an abnormal group will be present, but only approximately eleven (11%) percent of normals are present.
It is noted that a "Score" scale along the abscissa will be nonlinear, when compared to the (100-Specificity) scale.
FIG. 8 shows data presented in (ROC) format for Abnormals in various categories:
For subjects known to have had a myocardial infarction (MI) shown by twelve (12) lead (ECG), identified as (BEMI);
For subjects with non-specific ST-T wave abnormality on twelve (12) lead (ECG), identified as (BEST);
For a subjects with normal resting twelve (12) lead (ECG) but awaiting surgery, identified as (BNOB).
For a test set of patients who have (CAD), identified as (BTEST) and (BGENSIA).
FIG. 9 shows data plotted in FIG. 8 plotted in a different format in which the abcissa is scaled in terms of the "Score" developed by the present invention method.
Present is also a curve for Normals data, identified as (NORM).
Also included are curves for two groups additional groups of volunteer subjects which contain patients who have known risk factors for (CAD) identified as (BMAQ) and (BTNR). These constitute a "real-world" population of what are considered normals, in that both normals and abnormals are present. As would be expected, the data for the (BMAQ) and (BTNR) groups is generally positioned between the data for the known abnormal (BTEST) and normal groups.
The important thing to note is that the method of the present invention very definitely separates the various groups whether presented in the format of FIG. 8 or FIG. 9.
It is to be understood that throughout this Disclosure the RMS Mean values are cited. It is possible to utilize other caluclated values, such as Averages, in the method of the present invention. The term "Mean" should be interpreted broadly to include such alternatives.
The terms "Assumed Gaussian" have also been used throughtout this Disclosure when refering to Data Distribution RMS Means and RMS Standard Deviations. It is noted that in fact, analysis of empirically obtained data has proven the assumption to be valid.
To provide full disclosure a print-out of major portions of the computer program utilized in the practice of the present invention is included herewith directly. Also included following the computer print-out is a table of data which documents the above discussed results.
Note It has been found that the Score values corresponding to the Frank X lead 150-250 Hz catagory, (P5 in the text of the Disclosure), and the Z lead 10-60 Hz catagory, (P13 in the text of the Disclosure), are often not required in practice of the present invention as little additional information is provided thereby. The following computer program therefore utilizes only twenty-eight (28) Score Components in arriving at a Score. This demonstrated use is not to be interpreted as a limitation on the present invention. ##SPC1##
Finally, FIG. 10 provides a Flow Chart representation of the primary focus of the preferred embodiment of the Method of the present invention, said method comprising a noninvasive approach to investigating cardiac status of a subject, and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom.
The first steps (A) and (A') respectively, are shown to involve:
a. in step (A) obtaining data corresponding to (ECG) cycle(s) from at least one monitored lead(s) of an (ECG) system for a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality; and
b. in step (A') obtaining data corresponding to (ECG) cycle(s) from at least one monitored lead(s) of an (ECG) system for a subject, said monitored (ECG) system lead(s) utilized being the same as the monitored (ECG) system lead(s) utilized to obtain (ECG) cycle(s) for a multiplicity of normal subjects.
Next, in steps (B) and (B') respectively, a portion of an (ECG) cycle is selected, (e.g. while the QRS complex is preferred, any portion, or the entire (ECG) cycle can be selected), for each of the normal subject population (B) and subject (B'). The same (ECG) cycle portion is typically selected for both the normal Subject population and the subject.
Next, in steps (C) and (C'), in conjunction with application of filtering techniques, a plurality of data sets are arrived at for each monitored (ECG) system lead for both the normal subject population, (step (C)), and for the subject, (step (C')). Each of said data sets corresponds to a composite of said selected (ECG) cycle portion for said population of normal subjects in a selected frequency band range. It is disclosed that the preferred filtering technique is digital and has been utilized to provide data sets for:
a. data contained in all frequencies;
b. data contained in the frequency band of (0.0) to (10) HZ;
c. data contained in the frequency band of (10) to (60) HZ;
d. data contained in the frequency band of (60) to (150) HZ;
e. data contained in the frequency band of (150) to (250) HZ.
It is specifically disclosed, however, that data in only one frequency band range, (e.g. all frequencies), can be provided in this step. It is also noted that other frequency bands can be selected, and that the recited bands are exemplary and nonlimiting.
(It is noted that steps (A), (B), (C) and (E) might be carried out only once for many runs of steps (A'), (B'), (C') and (E'). This is because steps (A), (B), (C) and (E) are applied to a normal subject population, which does not change, except perhaps when additional normal subject data are added to a normal subject populating data bank. Steps (A'), (B'), (C') and (E') must be run anew for each subject tested, however. Steps (D) and (F) will therefore access relatively standard values for the normal subject population, while accessing new values for each subject tested. However, steps (A), (B), (C) and (E) are inherently performed in the context of any testing of a subject.)
In step (D), normal subject population and subject (ECG) data set(s), (formed by user determined filtering techniques, (steps (C) and (C')), are plotted and displayed as a function of at least one parameter selected from the group consisting of time and frequency. It is noted that where data is plotted as a function of frequency a time to frequency domain conversion calculation must be performed to provide the frequency domain data. The results of this step are to provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density. (It is noted that variations of the present invention procedure provide that, this step can omitted, or be performed after steps which directly follow, (e.g. steps (E) and (E') or step (F)). Step (D) is presented at this point in the flow chart only because the data to be plotted and displayed is available at this point. It is further noted that step (D) can be performed utilizing a different selected (ECG) cycle portion, in steps (B) and (B'), than is selected and utilized in following steps (E), (E') and (F)), (e.g. a full (ECG) cycle can be chosen for plotting and a QRS complex (ECG) cycle portion chosen for use in steps (E), (E') and (F)).
The next steps, (E) and (E') respectively, involve calculating corresponding representative parameter(s), and desired ratios thereof, from resulting data sets in said selected frequency band ranges for each monitored (ECG) system lead, for both the normal subject population (step (E)) and said subject, (step (E')). These steps are indicated as performed parallel to step (D) as data to allow both performance of step (D) and steps (E) and (E') is available at this point. (As noted above, the selected (ECG) cycle portion chosen in steps (B) and (B') and utilized in steps (E), (E') and (F) can be different than that utilized in step (D). It is also to be understood that in one version of the present invention procedure, steps (E), (E') and (F) can be omitted and only step (D) performed).
In step (F) corresponding subject and normal subject population representative parameter(s) and/or corresponding ratios of subject and ratios of normal subject population representative parameters are then compared and results of said comparison are combined to arrive at a "score", the magnitude of which "score" provides an indication of the cardiac status of said subject, and enables classification of a subject into normal and abnormal cardiac categories. A confidence level "acceptance test" can be optionally applied to qualify results of said comparisons for inclusion in arriving at said "score".
It is also noted that the step of calculating representative parameter(s) for normal subject population and subject data sets for monitored (ECG) system leads typically involves calculating at least one parameter selected from the group consisting of a root-mean-square mean and a root-mean-square standard deviation from said data set(s) from which composite(s) of a selected (ECG) cycle portion are calculated.
It is further noted that obtaining mean and standard deviation parameters, (typically, but not necessarily, based upon root-mean-square calculated parameters), enables practice of an "acceptance test" wherein a result of comparing subject to corresponding normal subject population parameters or ratios of a subject to corresponding ratios of the normal subject population parameters is accepted only if a subject acceptance parameter is offset by greater than, for instance, at least one normal subject population standard deviation from a mean of said normal subject population.
This step includes displaying of the "score" and when desired, components obtained from various composite data sets combined to arrive thereat.
Having hereby disclosed the subject matter of the present invention, it should be obvious that many modifications, substitutions and variations of the present invention are possible in light of the teachings. It is therefore to be understood that the invention may be practiced other than as specifically described, and should be limited in breadth and scope only by the Claims.
Claims
I claim:
1. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining data from (ECG) cycle(s) from each of a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality, by monitoring at least one lead(s) of an (ECG) system; b. selecting some (ECG) cycle portion, and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s) by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from each of a number of members of said multiplicity of members of a population of subjects who have been documented as normal subjects, each said calculated average selected (ECG) cycle portion data set being a composite data set of said selected (ECG) cycle portion for said population of normal subjects, for a monitored (ECG) system lead; c. obtaining data from (ECG) cycle(s) from a subject, by monitoring at least one lead(s) of an (ECG) system, said (ECG) system lead(s) monitored being the same as the monitored (ECG) system lead(s) utilized in step a. to obtain data utilized in step b.; d. selecting some (ECG) cycle portion, which is the same as that selected in step b., and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s) by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from said subject, each said calculated average selected (ECG) cycle portion data set being a composite data set of said selected (ECG) cycle portion for said subject, for a monitored (ECG) system lead; e. calculating corresponding representative parameter(s) from resulting composite data sets calculated in steps b. and d., for monitored (ECG) system lead(s), for, respectively, said normal subject population and said subject; f. comparing subject to corresponding normal subject population representative parameter(s), and combining results thereof to arrive at a "score", the magnitude of which "score" results from difference(s) between magnitude(s) of corresponding normal subject population, and subject representative parameter(s), which "score" magnitude increases when said difference(s) in magnitude(s) between corresponding normal subject population, and subject, representative parameter(s) increase, the magnitude of which "score" provides an indication of the cardiac status of said subject, with a "score" near zero being indicative of a subject properly categorized as a cardiac normal in that the magnitude(s) of subject representative parameter(s) are generally more closely matched to the magnitude(s) of corresponding normal subject population representative parameter(s), and with a progressively higher "score" being indicative of a subject progressively more properly categorized as a cardiac abnormal in that the magnitude(s) of subject representative parameter(s) are generally progressively less closely matched to the magnitude(s) of corresponding normal subject population representative parameter(s).
2. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 1, which further comprises performance of the steps of calculating, and comparing ratio(s) of subject, to corresponding ratio(s) of normal subject population, representative parameters, and combining results thereof with those from comparing subject to corresponding normal subject population, representative parameter(s), in arriving at said "score", the magnitude of which "score" then further results from difference(s) between magnitude(s) of corresponding normal subject population and subject ratio(s) of representative parameters, which "score" magnitude increases when difference(s) in magnitude(s) between corresponding normal subject population, and subject, ratio(s) of representative parameters increase.
3. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 2, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing specific ratio(s) of subject, to corresponding specific ratio(s) of normal subject population representative parameters prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining mean and standard deviation acceptance parameters for specific ratio(s) of normal subject population representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each corresponding specific ratio of subject representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific ratio of subject representative parameters, to a corresponding specific ratio of representative parameters for said normal subject population, in arriving at said "score", only if said acceptance parameter for said specific ratio of said subject representative parameters is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of said corresponding specific ratio of normal subject population representative parameters.
4. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 1, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing specific subject, to corresponding specific normal subject population, representative parameter(s) prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining mean and standard deviation acceptance parameters for specific normal subject population representative parameter(s) based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each corresponding specific subject representative parameter based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific subject representative parameter, to a corresponding specific normal subject population representative parameter, in arriving at said "score", only if the acceptance parameter for said specific subject representative parameter is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of said corresponding specific normal subject population representative parameter.
5. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 1, which further comprises performance of the steps of: a. providing at least one coordinate system selected from the group consisting of magnitude vs. time, and magnitude vs. frequency; b. for an (ECG) cycle portion, performing calculations necessary to plot and display and plotting and displaying normal subject population and subject (ECG) data as a function of at least one parameter selected from the group consisting of time and frequency, to respectively provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density data, observation of which provides an indication of the cardiac status of said subject; and c. observing and interpreting the displayed plot(s).
6. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 1, in which the step of comparing subject, to normal subject population, representative parameter(s) and combining the results thereof to arrive at a "score", includes performance of the step of selecting parameters from the group consisting of root-mean-square mean and root-mean-square standard deviation as representative parameters.
7. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining data from (ECG) cycle(s) from each of a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality, by monitoring at least one leads) of an (ECG) system; b. selecting some (ECG) cycle portion, and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s) by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from each of a number of members of said multiplicity of members of a population of subjects who have been documented as normal subjects, each said calculated average selected (ECG) cycle portion data set being a composite data set of said selected (ECG) cycle portion for said population of normal subjects, for a monitored (ECG) system lead; c. obtaining data from (ECG) cycle(s) from a subject, by monitoring at least one lead(s) of an (ECG) system, said (ECG) system lead(s) monitored being the same as the monitored (ECG) system lead(s) utilized in step a. to obtain data utilized in step b.; d. selecting some (ECG) cycle portion, which is the same as that selected in step b., and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s) by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from said subject, each said calculated average selected (ECG) cycle portion data set being a composite data set of said selected (ECG) cycle portion for said subject, for a monitored (ECG) system lead; e. calculating corresponding representative parameter(s) and corresponding ratio(s) involving representative parameters from resulting composite data sets calculated in steps b. and d., for monitored (ECG) system lead(s), for, respectively, said normal subject population and said subject; f. comparing specific ratio(s) of subject to corresponding specific ratio(s) of normal subject population representative parameters, and combining results thereof to arrive at a "score", the magnitude of which "score" results from difference(s) between magnitude(s) of specific corresponding ratio(s) of normal subject population, and ratio(s) of subject representative parameters, which "score" magnitude increases when said difference(s) in magnitude(s) between specific ratio(s) of corresponding normal subject population, and specific ratio(s) of subject representative parameters increase, the magnitude of which "score" provides an indication of the cardiac status of said subject, with a "score" near zero being indicative of a subject properly categorized as a cardiac normal in that the magnitude(s) of ratio(s) of subject representative parameters are generally more closely matched to the magnitude(s) of corresponding ratio(s) of normal subject population representative parameters, and with a progressively higher "score" being indicative of a subject progressively more properly categorized as a cardiac abnormal in that the magnitude(s) of ratio(s) of subject representative parameters are generally progressively less closely matched to the magnitude(s) of ratio(s) of corresponding normal subject population representative parameters.
8. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 7, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing ratio(s) of subject, to corresponding ratio(s) of normal subject population, representative parameters prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining acceptance mean and standard deviation parameters for specific ratio(s) of normal subject population representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each corresponding specific ratio of subject representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific ratio of subject representative parameters, to a corresponding specific ratio of representative parameters for said normal subject population, in arriving at said "score", only if the acceptance parameter for said specific ratio of said subject representative parameters is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of said corresponding specific ratio of normal subject population representative parameters.
9. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 7, which further comprises performance of the steps of: a. providing at least one coordinate system selected from the group consisting of magnitude vs. time, and magnitude vs. frequency; b. for an (ECG) cycle portion, performing calculations necessary to plot and display and plotting and displaying normal subject population and subject (ECG) data, as a function of at least one parameter selected from the group consisting of time and frequency, to respectively provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density data, observation of which provides an indication of the cardiac status of said subject; and c. observing and interpreting displayed plot(s).
10. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 7, in which the step of comparing specific ratio(s) of subject, to corresponding specific ratio(s) of normal subject population, representative parameter(s) and combining the results thereof to arrive at a "score", which "score" provides an indication of the presence or absence of cardiac abnormality in said subject, includes performance of the step of selecting parameters from the group consisting of root-mean-square mean and root-mean-square standard deviation as representative parameters.
11. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining data from (ECG) cycle(s) from each of a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality, by monitoring at least one lead(s) of an (ECG) system; b. selecting some (ECG) cycle portion and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from each of a number of said multiplicity of members of a population of subjects who have been documented as normal subjects, and selecting a plurality of frequency bands and applying filtering techniques, to provide a plurality of data sets for each said at least one (ECG) system lead(s) monitored, each said data set being a composite data set of said selected (ECG) cycle portion for said population of normal subjects in a monitored lead and selected frequency band range; c. obtaining data from (ECG) cycle(s) from a subject, by monitoring at least one lead(s) of an (ECG) system, said (ECG) system lead(s) monitored being the same as the monitored (ECG) system lead(s) utilized in step a. to obtain data utilized in step b.; d. selecting some (ECG) cycle portion, said (ECG) cycle portion being the same as that selected in step b. for said normal subject population, and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for subject (ECG) cycle(s), and selecting a plurality of frequency bands, said selected frequency bands being the same as those selected in step b. for said normal subject population, and applying filtering techniques which are the same as those applied in step b. for said normal subject population, to provide a plurality of data sets for each said at least one monitored (ECG) system lead(s), each said data set being a composite data set of said selected (ECG) cycle portion for said subject in a monitored lead and selected frequency band range; e. calculating corresponding representative parameter(s) from resulting composite data sets calculated in steps b. and d., in said selected frequency band ranges for monitored (ECG) system lead(s), for respectively, said normal subject population and said subject; f. comparing specific subject to specific normal subject population corresponding representative parameter(s), and combining results thereof to arrive at a "score", the magnitude of which "score" results from difference(s) between magnitude(s) of corresponding normal subject population and subject representative parameter(s), which "score" magnitude increases when said difference(s) in magnitude(s) between corresponding normal subject population and subject representative parameter(s) increase, the magnitude of which "score" provides an indication of the cardiac status of said subject, with a "score" near zero being indicative of a subject properly categorized as a cardiac normal in that the magnitude(s) of subject representative parameter(s) are generally more closely matched to the magnitude(s) of corresponding normal subject population representative parameter(s), and with a progressively higher "score" being indicative of a subject progressively more properly categorized as a cardiac abnormal in that the magnitude(s) of subject representative parameter(s) are generally progressively less closely matched to the magnitude(s) of corresponding normal subject population representative parameter(s).
12. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 11, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing specific subject, to corresponding specific normal subject population, representative parameters prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining mean and standard deviation acceptance parameters for specific normal subject population representative parameter(s) based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each corresponding specific subject representative parameter based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific subject representative parameter to a corresponding specific normal subject population representative parameter, in arriving at said "score", only if the acceptance parameter for said specific subject representative parameter is set off by at least one associated normal population acceptance standard deviation from the acceptance mean of said corresponding specific normal subject population representative parameter.
13. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 11, which further comprises performance of the steps of: a. providing at least one coordinate system selected from the group consisting of magnitude vs. time, and magnitude vs. frequency; b. for an (ECG) cycle portion, performing calculations necessary to plot and display and plotting and displaying normal subject population and subject (ECG) data in at least one frequency band range, as a function of at least one parameter selected from the group consisting of time and frequency, to respectively provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density data, observation of which provides an indication of the cardiac status of said subject; and c. observing and interpreting displayed plot(s).
14. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 11, in which the step of comparing corresponding subject, to normal subject population, representative parameter(s) and combining the results thereof to arrive at a "score", includes performance of the step of selecting parameters from the group consisting of root-mean-square mean and root-mean-square standard deviation.
15. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining data from (ECG) cycle(s) from each of a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality, by monitoring at least one lead(s) of an (ECG) system; b. selecting some (ECG) cycle portion and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from each of a number of said multiplicity of members of a population of subjects who have been documented as normal subjects, and selecting a plurality of frequency bands and applying filtering techniques, to provide a plurality of data sets for said at least one (ECG) system lead(s) monitored, each said data set being a composite data set of said selected (ECG) cycle portion for said population of normal subjects in a monitored lead and selected frequency band range; c. obtaining data from (ECG) cycle(s) from a subject, by monitoring at least one lead(s) of an (ECG) system, said (ECG) system lead(s) monitored being the same as the monitored (ECG) system lead(s) utilized in step a. to obtain data utilized in step b.; d. selecting some (ECG) cycle portion, said (ECG) cycle portion being the same as that selected in step b. for said normal subject population, and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for subject (ECG) cycle(s), and selecting a plurality of frequency bands, said selected frequency bands being the same as those selected in step b. for said normal subject population, and applying filtering techniques which are the same as those applied in step b. for said normal subject population, to provide a plurality of data sets for said at least one monitored (ECG) system lead(s), each said data set being a composite data set of said selected (ECG) cycle portion for said subject in a monitored lead and selected frequency band range; e. calculating corresponding representative parameter(s) and corresponding ratio(s) involving representative parameters from resulting composite data sets calculated in steps b. and d., in said selected frequency band ranges for monitored (ECG) system lead(s), for respectively, said normal subject population and said subject; f. comparing specific ratio(s) of subject to corresponding specific ratio(s) of normal subject population representative parameters, and combining results thereof to arrive at a "score", the magnitude of which "score" results from difference(s) between magnitude(s) of corresponding specific ratio(s) of normal subject population and specific ratio(s) of subject representative parameters, which "score" magnitude increases when said difference(s) in magnitude(s) between specific ratio(s) of corresponding normal subject population and subject representative parameters increase, the magnitude of which "score" provides an indication of the cardiac status of said subject, with a "score" near zero being indicative of a subject properly categorized as a cardiac normal in that the magnitude(s) Of ratio(s) of subject representative parameters are generally more closely matched to the magnitude(s) of corresponding ratio(s) of normal subject population representative parameters, and with a progressively higher "score" being indicative of a subject progressively more properly categorized as a cardiac abnormal in that the magnitude(s) of ratio(s) of subject representative parameters are generally progressively less closely matched to the magnitude(s) of ratio(s) of corresponding normal subject population representative parameters.
16. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 15, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing specific ratio(s) of subject, to corresponding specific ratio(s) of normal subject population, representative parameters prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining mean and standard deviation acceptance parameters for specific ratio(s) of normal subject population representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each corresponding specific ratio of subject representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific ratio of subject representative parameters, to a corresponding specific ratio of representative parameters for said normal subject population, in arriving at said "score", only if the acceptance parameter for said specific ratio of said subject representative parameters is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of said specific ratio of normal subject population representative parameters.
17. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 15, which further comprises performance of the steps of: a. providing at least one coordinate system selected from the group consisting of magnitude vs. time, and magnitude vs. frequency; b. for an (ECG) cycle portion, performing calculations necessary to plot and display and plotting and displaying normal subject population and subject (ECG) data for at least one frequency band range, as a function of at least one parameter selected from the group consisting of time and frequency, to respectively provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density data, observation of which provides an indication of the cardiac status of said subject; and c. observing and interpreting displayed plot(s).
18. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 15, in which the step of comparing ratio(s) of subject, to corresponding ratio(s) of normal subject population, representative parameters and combining the results thereof to arrive at a "score", includes performance of the step of selecting ratios of parameter(s) from the group consisting of a ratio of a root-mean-square mean to at least one other root-mean-square mean, and a ratio of at least one root-mean-square mean to a root-mean-square standard deviation.
19. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining data from (ECG) cycle(s) from each of a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality, by monitoring at least one lead(s) of an (ECG) system; b. selecting some (ECG) cycle portion and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from each of a number of said multiplicity of members of a population of subjects who have been documented as normal subjects, and selecting a plurality of frequency bands and applying filtering techniques, to provide a plurality of data sets for said at least one (ECG) system lead(s) monitored, each said data set being a composite data set of said selected (ECG) cycle portion for said population of normal subjects in a monitored lead and selected frequency band range; c. obtaining data from (ECG) cycle(s) from a subject, by monitoring at least one lead(s) of an (ECG) system, said (ECG) system lead(s) monitored being the same as the monitored (ECG) system lead(s) utilized in step a. to obtain data utilized in step b.; d. selecting some (ECG) cycle portion, said (ECG) cycle portion being the same as that selected in step b. for said normal subject population, and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for subject (ECG) cycle(s), and selecting a plurality of frequency bands, said selected frequency bands being the same as those selected in step b. for said normal subject population, and applying filtering techniques which are the same as those applied in step b. for said normal subject population, to provide a plurality of data sets for said at least one monitored (ECG) system lead(s), each said data set being a composite data set of said selected (ECG) cycle portion for said subject in a monitored lead and selected frequency band range; calculating corresponding representative parameter(s) and corresponding ratio(s) involving representative parameters from resulting composite data sets calculated in steps b. and d., in said selected frequency band ranges for monitored (ECG) system lead(s), for respectively, said normal subject population and said subject; f. comparing specific subject and corresponding specific normal subject population representative parameter(s), and combining results thereof with the results of comparing specific ration(s) of subject to corresponding specific ratio(s) of normal subject population representative parameters, to arrive at a "score", the magnitude of which "score" results from difference(s) in magnitude(s) between corresponding subject and normal subject population representative parameter(s) and difference(s) between magnitude(s) of corresponding ratio(s) of normal subject population, and ratio(s) of subject representative parameters, which "score" magnitude increases when difference(s) in magnitude(s) between corresponding subject and normal subject population representative parameter(s) increase and difference(s) in magnitude(s) between ratio(s) of corresponding normal subject population, and ratio(s) of subject representative parameters increase, the magnitude of which "score" provides an indication of the cardiac status of said subject, with a "score" near zero being indicative of a subject properly categorized as a cardiac normal in that magnitude(s) of subject representative parameter(s) are generally more closely matched to the magnitude(s) of corresponding normal subject population representative parameter(s) and magnitude(s) of ratio(s) of subject representative parameters are generally more closely matched to the magnitude(s) of corresponding ratio(s) of normal subject population representative parameters, and with a progressively higher "score" being indicative of a subject progressively more properly categorized as a cardiac abnormal in that magnitude(s) of subject representative parameter(s) are generally progressively less closely matched to the magnitude(s) of corresponding normal subject population representative parameter(s) and magnitude(s),of ratio(s) of subject representative parameter(s) are generally progressively less closely matched to the magnitude(s) of ratio(s) of corresponding normal subject population representative parameters.
20. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 19, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing subject, to corresponding normal subject population representative parameter(s), and the results of comparing ratios of subject representative parameters to corresponding ratios of normal subject population representative parameters, prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining mean and standard deviation acceptance parameters for specific normal subject population representative parameter(s) and specific ratio(s) of normal subject population representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each corresponding specific subject representative parameter and each corresponding specific ratio of subject representative parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific subject representative parameter to a corresponding specific normal subject population representative parameter in arriving at said "score", only if the acceptance parameter for said specific subject representative parameter is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of the corresponding specific normal subject population representative parameter; and accepting the results of comparing a specific ratio of subject representative parameters to a corresponding specific ratio of representative parameters for said normal subject population, in arriving at said "score", only if the acceptance parameter of said specific ratio of said subject representative parameters is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of said corresponding specific ratio of normal subject population representative parameters.
21. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 19, which further comprises performance of the steps of: a. providing at least one coordinate system selected from the group consisting of magnitude vs. time, and magnitude vs. frequency; b. for an (ECG) cycle portion, performing calculations necessary to plot and display and plotting and displaying normal subject population and subject (ECG) data for at least one frequency band range, as a function of at least one parameter selected from the group consisting of time and frequency, to respectively provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density data, observation of which provides an indication of the cardiac status of said subject; and c. observing and interpreting displayed plot(s).
22. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 19, in which the step of comparing corresponding subject, to normal subject population, representative parameter(s) and ratio(s) thereof, and combining the results thereof to arrive at a "score", which "score" provides an indication of the presence or absence of cardiac abnormality in said subject, includes performance of the step of selecting representative parameter(s) and ratio(s) thereof formed between representative parameters from the group consisting of root-mean-square mean and root-mean-square standard deviation.
23. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining (ECG) data from (ECG) cycle(s) at each lead of a three lead Frank X-Y-Z (ECG) system, from each of a multiplicity of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality; b. selecting the QRS portion of said obtained (ECG) cycle(s) obtained from each Frank X-Y-Z (ECG) system lead and calculating an average selected QRS cycle portion data set for each monitored Frank (X-Y-Z) (ECG) system lead, by, for each Frank (X-Y-Z) (ECG) system lead, a procedure comprising combining corresponding QRS cycle portion data points for said selected QRS cycle portion for (ECG) cycle(s) obtained from each of a number of members of said multiplicity of members of a population of subjects who have been documented as normal subjects, and selecting a plurality of frequency bands and applying filtering techniques, to provide a plurality of data sets for each Frank X-Y-Z (ECG) system lead, each said data set being a composite QRS data set for said population of normal subjects, for a Frank (X-Y-Z) (ECG) system lead, in a selected frequency band range; c. obtaining (ECG) data from (ECG), cycle(s) from each Frank X-Y-Z (ECG) system lead for a subject; d. selecting the QRS complex of said obtained (ECG) cycle(s) obtained at each Frank X-Y-Z (ECG) system lead, said (ECG) QRS cycle portion being the same as that selected in step b. for said normal subject population, and calculating an average selected QRS cycle portion data set for each monitored Frank (X-Y-Z) (ECG) system lead, by, for each Frank (X-Y-Z) (ECG) system lead, a procedure comprising combining corresponding QRS cycle portion data points for said selected QRS cycle portion for subject (ECG) cycle(s), and selecting a plurality of frequency bands, said frequency bands being the same as those selected in step b. for said normal subject population and applying filtering techniques which are the same as those applied in step b. for said normal subject population, to provide a plurality of data sets for each Frank X-Y-Z (ECG) system lead, each said data set being a composite QRS data set for said subject, for a Frank (X-Y-Z) (ECG) system lead, in a selected frequency band range; e. calculating corresponding root-mean-square mean and root-mean-square standard deviation parameter(s) as well as corresponding ratio(s) involving root-mean-square and root-mean-square standard deviation parameters from resulting composite QRS data sets calculated in steps b. and d., in each selected frequency band range for each Frank X-Y-Z (ECG) system lead, for, respectively, said normal subject population and subject; f. comparing specific subject and corresponding specific normal subject population root-mean-square parameter(s), and combining results thereof with the results of comparing specific ratio(s) of subject to corresponding specific ratio(s) of normal subject population root-mean-square parameters, to arrive at a "score", the magnitude of which "score" results from difference(s) in magnitude(s) between corresponding subject and normal subject population root-mean-square parameter(s) and difference(s) between magnitude(s) of corresponding ratio(s) of normal subject population, and ratio(s) of subject root-mean-square parameters, which "score" magnitude increases when difference(s) in magnitude(s) between corresponding subject and normal subject population root-mean-square parameter(s) increase and difference(s) in magnitude(s) between ratio(s) of corresponding normal subject population, and ratio(s) of subject root-mean-square parameters increase, the magnitude of which "score" provides an indication of the cardiac status of said subject, with a "score" near zero being indicative of a subject properly categorized as a cardiac normal in that magnitude(s) of subject root-mean-square parameter(s) are generally more closely matched to the magnitude(s) of corresponding normal subject population root-mean-square parameter(s) and magnitude(s) of ratio(s) of subject root-mean-square parameters are generally more closely matched to the magnitude(s) of corresponding ratio(s) of normal subject population root-mean-square parameters, and with a progressively higher "score" being indicative of a subject progressively more properly categorized as a cardiac abnormal in that magnitude(s) of subject root-mean-square parameter(s) are generally progressively less closely matched to the magnitude(s) of corresponding normal subject population root-mean-square parameter(s) and magnitude(s) of ratio(s) of subject root-mean-square parameter(s) are generally progressively less closely matched to the magnitude(s) of ratio(s) of corresponding normal subject population root-mean-square parameters.
24. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 23, which further comprises performance of the step of applying a confidence level acceptance test to results of comparing corresponding specific normal subject population to specific subject root-mean-square parameters, and corresponding specific ratios of said root-mean-square parameters of subject, to specific ratios of normal subject population, root-mean-square parameters prior to combining results thereof to arrive at a "score", said confidence level test comprising: a. the step of determining mean and standard deviation acceptance parameters for specific normal subject population root-mean-square parameter(s) and specific ratio(s) of normal subject population root-mean-square parameter(s) based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said population of normal subjects was calculated; b. the step of determining an acceptance parameter for each specific corresponding subject root-mean-square parameter and each corresponding specific ratio of subject root-mean-square parameters based upon the (ECG) data from which said composite data set of said selected (ECG) cycle portion for said subject was calculated; and c. the step of accepting the results of comparing a specific subject root-mean-square parameter to a corresponding specific normal subject population root-mean-square parameter in arriving at said "score", only if the acceptance parameter for said specific subject root-mean-square parameter is set off by at least one associated normal subject population acceptance standard deviation from the acceptance mean of the corresponding specific normal subject population root-mean-square parameter; and accepting the results of comparing a specific ratio of subject root-mean-square parameters to a corresponding specific ratio of root-mean-square parameters for said normal subject population, in arriving at said "score", only if the acceptance parameter of said specific ratio of said subject root-mean-square parameters is set off by at least one associated normal population acceptance standard deviation from the acceptance mean of said corresponding specific ratio of normal subject population root-mean-square parameters.
25. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 24, in which the step applying a confidence level test for accepting differences between corresponding normal subject population to subject root-mean-square parameters, and corresponding ratios of said root-mean-square parameters, for each Frank X-Y-Z (ECG) system lead and using said accepted differences to arrive at a "score", includes performing the step of determining if an acceptance parameter of the subject falls outside a region of distribution of corresponding normal subject population data, centered around an acceptance mean for said corresponding normal subject population, which region of distribution of normal subject population data includes ninety-five (95%) percent of said corresponding normal subject population distribution data.
26. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 23, which includes in performance of the step of applying filtering techniques in steps b. and d. thereof, the specific formation of fifteen (15) data sets for each of the normal subject population and the subject, said data sets being five (5) per each Frank X-Y-Z (ECG) system lead, said five (5) data sets per each Frank X-Y-Z (ECG) system lead covering: a. all frequencies present; b. the frequency range of zero (0) to ten (10) Hertz; c. the frequency range of ten (10) to sixty (60) Hertz; d. the frequency range of sixty (60) to one-hundred-fifty (150) Hertz; and e. the frequency range of one-hundred-fifty (150) to two-hundred-fifty (250) Hertz.
27. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 23, which further comprises performance of the steps of: a. providing at least one coordinate system selected from the group consisting of magnitude vs. time, and magnitude vs. frequency; b. for an (ECG) cycle portion, performing calculations necessary to plot and display and plotting and displaying normal subject population and subject (ECG) data for at least one frequency band range, as a function of at least one parameter selected from the group consisting of time and frequency, to respectively provide as desired, visually interpretable plots of (ECG) magnitude and power spectral density data, observation of which provides an indication of the cardiac status of said subject; and c. observing and interpreting displayed plot(s).
28. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 23, in which corresponding ratios of said root-mean-square parameters are compared, said method including performance of the steps of: a. selecting said ratios to be calculated as, for each (ECG) lead and frequency band range utilized, a root-mean-square mean for said frequency band range divided by a sum of root-mean-square means of all frequency band ranges present; said ratio for a normal subject population being subtracted from a corresponding ratio for said subject, each said subtracted difference being divided by a root-mean-square standard deviation for said normal subject population in a corresponding frequency band range; and b. performing the calculations described in step a.
29. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom as in claim 23, in which corresponding ratios of said root-mean-square parameters are compared, includes performance of the steps of: a. selecting said ratios to be calculated as, for each (ECG) lead and frequency band range utilized, a root-mean-square mean for one (ECG) lead divided by a root-mean-square mean for another (ECG) lead; said ratio for a normal subject population being subtracted from a corresponding ratio for said subject, each said subtracted difference being divided by a root-mean-square standard deviation for said normal subject population ratio in a corresponding frequency band range; and b. performing the calculations described in step a.
30. A noninvasive method of investigating cardiac status of a subject and enabling classification of a subject into normal and abnormal cardiac categories utilizing electrocardiography (ECG) data obtained therefrom, said method comprising, in a functional sequence, performance of the steps of: a. obtaining data from (ECG) cycle(s) from each of a multiplicity of members of a population of subjects who have been documented as normal subjects, in that they do not show risk factors for, or demonstrate detectable cardiac abnormality, by monitoring at least one lead(s) of an (ECG) system; b. selecting some (ECG) cycle portion and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for (ECG) cycle(s) obtained from each of a number of said multiplicity of members of a population of subjects who have been documented as normal subjects, and selecting a plurality of frequency bands and applying filtering techniques, to provide a plurality of data sets for said at least one (ECG) system lead(s) monitored, each said data set being a composite data set of said selected (ECG) cycle portion for said population of normal subjects in a monitored lead and selected frequency band range; c. obtaining data from (ECG) cycle(s) from a subject, by monitoring at least one lead(s) of an (ECG) system, said (ECG) system lead(s) monitored being the same as the monitored (ECG) system lead(s) utilized in step a. to obtain data utilized in step b.; d. selecting some (ECG) cycle portion, said (ECG) cycle portion being the same as that selected in step b. for said normal subject population, and calculating an average selected (ECG) cycle portion data set for at least one monitored (ECG) system lead(s), by, for a monitored (ECG) system lead, a procedure comprising combining corresponding (ECG) cycle portion data points for said selected (ECG) cycle portion for subject (ECG) cycle(s), and selecting a plurality of frequency bands, said selected frequency bands being the same as those selected in step b. for said normal subject population, and applying filtering techniques which are the same as those applied in step b. for said normal subject population, to provide a purality of data sets for said at least one monitored (ECG) system lead(s), each said data set being a composite data set of said selected (ECG) cycle portion for said subject in a monitored lead and selected frequency band range; e. performing at least one of the following steps f. and g.; f. calculating mean and standard deviation representative parameters from at least one resulting composite data set calculated in step b., said mean and standard deviation parameters being from a selected frequency band range for a monitored (ECG) system lead, and providing a coordinate system consisting of magnitude vs. time on ordinate and abscissa respectively and plotting and displaying on said coordinate system loci consisting of: 1. normal subject population standard deviation bounds located above and below said normal subject population mean, and 2. a corresponding subject data set, then observing differences between normal subject population and subject data plots, with a subject data set falling within the normal subject standard deviation bounds being indicative of a subject properly classified as a cardiac normal, and with a subject data set falling progressively further outside said normal subject standard deviation bounds being indicative of a subject progressively more properly classified as a cardiac abnormal; g. calculating power spectral density data for at least one resulting composite data set calculated in step b. and for a corresponding resulting composite data set calculated in step d. for a selected frequency band range for a monitored (ECG) system lead, and providing a coordinate system consisting of magnitude vs. frequency on ordinate and abscissa respectively and plotting and displaying on said coordinate system power spectral density data loci, then observing differences between normal subject population and subject data loci, with closely matched corresponding subject and normal subject population data set power spectral density loci being indicative of a subject properly classified as a cardiac normal and with progressively more mismatched subject and normal subject population data set power spectral density loci being indicative of a subject progressively more properly classified as a cardiac abnormal.
Patent Citations (12)
| Patent | Date | Inventor | Cited By |
|---|---|---|---|
| US4667682 | 1987-05-01 | Ihlenfeld, III | |
| US4680708 | 1987-07-01 | Ambos | |
| US4974162 | 1990-11-01 | Siegel et al. | |
| US5020540 | 1991-06-01 | Chamoun | |
| US5025794 | 1991-06-01 | Albert et al. | |
| US5027824 | 1991-07-01 | Dougherty et al. | |
| US5046504 | 1991-09-01 | Albert et al. | |
| US5077667 | 1991-12-01 | Brown et al. | |
| US5092341 | 1992-03-01 | Kelen | |
| US5215099 | 1993-06-01 | Haberl et al. | |
| US5257621 | 1993-11-01 | Bardy et al. | |
| US5509425 | 1996-04-01 | Feng |
Non-Patent Literature (6)
- "Perioperative Cardiac Morbidity: Concepts & Co-Travoties" Mangano.
- "High Frequency QRS Electrocardiography in the Detection of Perfusion Following Thrombelytic Therapy" Sveriano et al, Clin. Cardio., 17, 1994.
- "Prognosis & Management AFter a First Myocardial Infartion", Noll & Benhorn, New England J. Med. vol. 323, No. 10, 1990.
- "The Meaning of the Area Under a Receiver-Operating Characteristic (ROC) Curve", Hawley & McNeil, Diagnostic Radiology, Apr. 1982.
- Predictive Valve of a Single Diagnostic Testing for Selected Population, Vecchio, New England J. Med., May 26, 1966.
- Phage and Group-Delay Characteristics of Signed-Averaged Electrocardiography for Patients with Ventricular Tachycardia IEEE Trans. in Bio Med, V42, No. 1, Jan. 1995.