DD281957A5 - REAL-TIME METHOD FOR COMPUTER-ASSESSED ARRHYTHMIA ANALYSIS DEVICES FOR DESCRIPTION OF THE STRUCTURE OF CARDIAC RHEIMTHMENTS - Google Patents

Abstract

The invention relates to a real-time method for computer-aided Arrhythmieanalysegeraete to describe the structure of heart rhythms. The object is achieved in that information about the time of occurrence of the R-waves or QRS complexes and the P-waves is continuously detected according to known signal processing methods and according to the invention in an event-controlled observation window of suitable length, preferably the last 8 heartbeats, the instantaneous heart rate is transformed into a logarithmic image area, a purposeful summary in heart rate classes, mean, range of variation, diversity and structure of the sequences of the heart rate classes are determined so that a complex description of the degree of regularity or irregularity in the observation window and a formulation of suspicions is made and, taking into account further features, a targeted leap into a branch of decision logic incorporating the appropriate ECG assessment. {arrhythmia analysis system; Electrocardiogram; Real-time process; Heart rates; Transformation; Class education; Structural description; regularity; Formulation of suspicions; Decision logic}

Description

Characteristic of the known state of the art

It is known that the system for ECG evaluation in real-time computer-aided arrhythmia analysis devices is controlled either by the recognized and classified QRS morphology, or that the QRS sequence is examined only or additionally as a point process (Feldmann, CL: Computer Detection of Cardiac Arrhythmias-Historic Review - IEEE Frontiers of Engineering and Computing in Health Care 1983 [5,1983, Columbus, Ohio]: Proceedings / IEEE Engineering in Medicine and Biology Society-1983).

The known methods are limited to the detection of electrical chamber activities and thus provide a relatively limited number of ECG diagnoses derived from the time series of QRS complexes and R-waves and QRS shape analysis. Furthermore, it is known that the range of possible ECG evaluations can be expanded by the inclusion of P wave detection for real-time systems to clinically significant ECG diagnoses that allows to take into account disturbances of electrical atrial activity and AV conduction (DD 263227).

The disadvantage is that in this case the decision logic greatly expanded and lack algorithms that effectively clarify the nature of the ECG history with respect to the diagnosis based on global criteria and rationally selects the right, further describing branches of the decision tree.

Object of the invention

The aim of the invention is a real-time method for computer-aided arrhythmia analysis systems for describing the structure of heart rhythms as a starting point for improved possibilities of real-time capable analysis of cardiac arrhythmias.

Explanation of the essence of the invention

The object of the invention is to provide a method that allows for the detection of R-waves and P-waves in the time regime of continuous signal processing a computationally favorable description of the regularity of a suitably long episode of recent heart actions and conclusions about the presence of cardiac arrhythmias eg strong sinus arrhythmia, extrasystole, AV conduction failures, absolute arrhythmia, etc.

The object is achieved in that according to known signal processing method information about the time of occurrence of the R-waves or QRS complexes and the P-waves are continuously detected and according to the invention in an event-controlled observation window of suitable length, preferably the last 8 heartbeats, the instantaneous heart rate is transformed into a logarithmic image area, an appropriate summary in heart rate classes, mean, range of variation, diversity and structure of the sequences of the heart rate classes are determined so that a complex description of the degree of regularity or irregularity in the observation window and a formulation of suspicions is made and taking into account further features, a targeted leap into a branch of decision logic containing the appropriate ECG assessment is made

 According to the invention, the transformation is achieved by an approximation of the logarithmic function

HF

HFKL = 16Id -r-- (1)

16bpm

reached by means of a stair curve with non-equidistant support points. It combines the advantages of the logarithmic image area with the advantages of a histogram with computationally and physiologically optimized class widths, in which the normal, non-pathological fluctuation range of the heart rate occupies only a few adjacent classes and fulfills the requirements for real-time capability

 The result of the approximation is the following table:

HFKL HF / bpm HFKL group

0 ... 15 16 (1) 31 1

16 ... 31 31 (2) 63 2

32 ... 47 64 (4) 127 3

48 ... 63 128 (8) 255 4

64 ... 79 256 (16) 511 5

In brackets the class widths are given in beat per minute (bpm).

In adaptation to special tasks (for example: examination of heartbeat sequences in humans or at the mouse) parts of the table or extensions according to the recognizable education law can be used. Also, the introduction of an open-top class (2.b., a collection class HFKL = 64 for all heart rates HF> 255bpm, e.g., for adult human examination) or extension of low heart rate exclusion (e.g.

HF <128 bpm instead of all HF <16 bpm, for example for mouse examinations) can be used to adapt to the respective task.

For the presentation of a heartbeat sequence in the image area, four essential features are described which, in conjunction with other known features, make the detection of cardiac arrhythmias possible in a computationally favorable manner.

These features for the detection of the indicated cardiac arrhythmias are determined in the image area of the heart rate for the heartbeats in the observation window and are

- The range of variation of the heart rate classes as the difference between the highest and lowest occurring heart rate class

- The variety of heart rate classes as the number of occurring, different heart rate classes

- an average of all HFKL values in the observation window

the structure of the sequences of the heart rate classes within the observation window (for the detection of bigeminy, trigemenia, salvos of extrasystoles, etc.) as a sequence of the intervals of the heart rate class of each individual beat to the heart rate class of a heartbeat with normal transition of the sinus node excitation (normal beat) or to another reference Heart rate class (eg the lowest, middle or highest, occupied heart rate class in the observation window) or as a sequence of characteristics derived from these intervals.

Together with further features that do not derive from the further processing of the instantaneous heart rate according to the invention and, in addition to statements on a QRS broadening, in particular the number of P waves in the observation window, their time reference to the QRS complexes and T waves, the regularity of the P values. Wave sequence within the observation window and the regularity of the time series of the smallest PQ intervals in the heartbeats, can be detected as cardiac arrhythmia, in addition to ventricular extrasystole, supraventricular extrasystole, ventricular rhythms, z. B. also include strong sinus arrhythmia, absolute arrhythmia and AV conduction disorders. According to the invention, the logic is based on the fact that, due to the regularity or special structure of the heartbeat sequence, different ECG evaluations are grouped together, for each group a subsystem of a decision tree is created which realizes the further differentiation into the final evaluation the global assessment of the regularity of the heartbeat sequence is jumped on immediately.

Thus, it is possible that with increasing variation width V and manifold I of the heartbeat sequence as global characteristics of regularity, the following groups are formed:

a) R waves regularly (V <V1) л (I <11):

- sinus rhythm

- AV block 2nd degree with fixed transfer ratio

- total AV block

- AV rhythm, ventricular rhythm

b) R-waves slightly irregular (V 1 SV <V2)

- strong sinus arrhythmia

- Disturbed heartbeat without recognizable extrasystoles and no detectable single AV conduction failures

c) R-waves strongly irregular and large manifold (V> V2) л (I> 12)

- massive polytopic extrasystole

- absolute arrhythmia

- AV block 2nd degree, type Wenckebach

d) R-serrations strongly irregular and low manifold (V> V2) л (I <12)

- nomotope extrasystole

- single AV bypass failures.

The further differentiation within a Arrhythmiebewertungsgruppe then takes place with the inclusion of the other features mentioned. V1, V2 as well as 11 and 12 are pre-learned threshold values from corresponding ECG data. The presented solution of the problem is even more expandable in arrhythmia analysis systems that use morphology differentiations for QRS complexes by grouping the various QRS morphologies and ordering them according to their distance to a reference morphology, preferably that of the normal QRS complex. The assessment of the breadth of variation, diversity and structure of the morphology class sequence in the observation window significantly increases the safety for the ECG final evaluation

 The advantages of the method are that

the natural fluctuation range of the instantaneous heart rate as a result of normal, non-pathological sinus arrhythmia is reflected in a few adjacent heart rate classes and can thus be easily eliminated from further consideration; it does not burden further processing

- Extrasystoles clearly with more than 7 heart rate categories above and AV conduction failures are significantly more than 15 heart rate classes below the heart rate class of a normal beat

- Absolute arrhythmia and massive polytopic extrasystoles can be effectively separated by means of the features variation width and diversity in the image area of the heart rate as well as number and type of P-waves in the observation window of the last heart beats

- The transformation into the logarithmic image area is computationally low feasible and thus a previously device technology not feasible significant rationalization effect in the handling of extensive decision logic for ECG evaluation in real-time operation is achieved.

The transformation into the heart rate classes can be ζ. ß. after the relationship

HFKL = int [(HF-HFGU) / HFKLB] + HFKLU (2)

HFGU = lowest heart rate of the class group

HFKLU = lowest heart rate class of the class group

respectively.

The required number of queries a for the group membership in the classification of the current heart rate HF to an HFKL group depends on the number Z of existing groups and is

a = 1 d Z,

or a = η for 2 ">Z> 2 ⁿ " ¹ .

According to equation (2), the transformation into the optimized logarithmic image area can thus be carried out with only a comparison operations, 1 addition, 1 subtraction and maximum (Z-1) bit shifts and is sufficiently fast for real-time operation.

embodiment

The invention is explained in more detail using the example of part of the decision logic for an arrhythmia analysis system. The accompanying drawings show in

Fig. 1: Overall sequence

Fig. 2: Update

Fig.3: Test for arrhythmia

Fig.4: Gross test on individual AV-Überleitungsausfall

Fig. 5: Gross test for individual extrasystole

Fig. 6: Fine test for individual extrasystole

Fig. 7: Fine test on individual AV-Überleitungsausfall

The part of the decision logic is implemented in an arrhythmia analysis system, which u. a. to identify absolute arrhythmia, single extrasystoles and single AV conduction failures.

Detection of such arrhythmias is an integral part of both cardiac monitoring and electrocardiological functional diagnostics.

In the exemplary embodiment, the identification of extrasystoles and AV conduction failures is not related to the heart rate class of a "normal heartbeat", ie, a cardiac action that normal conducts the atrial pacing to the ventricles.For practical reasons, an average of all heart rate classes in the observation window is used instead This leads to a change of the threshold values VAV and VEX in relation to the stated heart rate class intervals of 7 or 15 heart rate classes, whereby additionally the type of averaging has to be considered.

After the passage of a new RR interval, the corresponding features for the current heart action and the observation window are updated over the last 8 heart actions (FIG. 2), for the given task eliminating uninteresting, regular heartbeat sequences based on the variation width (V <2), absolute Arrhythmias on the basis of variation width V, diversity I, number PV or type of P-waves detected (Figure 3), a coarse test for individual AV conduction failures based on the variation width thresholds VAV (Figure 4) and a negative test based on a gross test of the variation width threshold value VEX to individual extrasystoles (FIG. 5). If a positive test result gives reason to suspect that a single AV conduction failure may have occurred (Figure A), a fine test (Figure 7) follows, in which an RR interval prolongation due to compensatory pause after a previous extrasystole is excluded and AV bypass failure is considered as confirmed only if special P-wave conditions are met. If a positive test result gives reason to suppose that the last heartbeat might be an extrasystole (Figure 5), a fine test follows, excluding the RR interval shortening due to a previous interposed extrasystole (Figure 6).

The tests for compensatory pause in (Figure 7) and interposed extrasystole in (Figure 6) are based on studies in the original range, so that the influence of normal biological variability z. B. due to normal sinus arrhythmia, must be considered by a range of fluctuation of the cardiac period to HPDA.

Claims (7)

1. A real-time method for computer-based Arrhythmieanalysegeräte to describe the structure of heart rhythms for detecting arrhythmia, which continuously detects information on the timing of the occurrence of R-waves or QRS complexes and the P-waves according to known signal processing methods and an instantaneous heart rate of all heartbeats limited heartbeat sequence determined from the RR interval, characterized in that - A transformation of the instantaneous heart rate in a logarithmic image area, which supports the next step of class formation, with subsequent Formation of ordered heart rate classes in the image area is performed, a description of the irregularity of the limited heartbeat sequence in the image area on the basis of the breadth of variation and number of occupied, different classes and - An assignment of the irregularity described to predetermined entry points for further processing in a decision logic based on pre-learned thresholds and below - The processing of individual, reduced decision trees according to medical standards is realized. 2. The method according to claim 1, characterized in that the transformation by the logarithmic function HFKL = 16Id ^HF 16bpm or their approximation is done. 3. The method according to claim 1 and 2, characterized in that the assignment of the heart rate to a heart rate class according to the table is realized. 4. The method according to claim 1 to 3, characterized in that the assignment of the heart rate to a heart rate class according to the relationship HFKL = int ([HF-HFGU] / HFKLB + HFKLU) is realized. 5. The method according to claim 1 to 4, characterized in that the decision logic is based on that due to the variability and number of occupied, different classes in the image area described regularity or irregularity of the heart rhythms and the structure of the sequence of classes and their mean different ECG evaluations are grouped together, a subsystem of a decision tree is created for each group, which realizes the further differentiation into the final scores, and which is jumped immediately as a result of the global assessment of the regularity and structure of the heartbeat sequence. 6. The method according to claim 1 to 5, characterized in that instead of or in addition to the ordered heart rate classes ordered morphology classes of the QRS complexes are used. For this 7 pages drawings Field of application of the invention Field of application of the invention is medicine. The invention relates to a method in the context of real-time processing of the electrocardiogram by means of digital computers, in particular microcomputers, for the analysis or monitoring of cardiac arrhythmias.