WO2023195856A1

WO2023195856A1 - Computer implemented method and system for ecg analysis

Info

Publication number: WO2023195856A1
Application number: PCT/NL2023/050187
Authority: WO
Inventors: VAN Peter Michael DAM; VAN Eelco Mattias DAM
Original assignee: Peacs B.V.; Peacs Investment B.V.
Priority date: 2022-04-07
Filing date: 2023-04-11
Publication date: 2023-10-12

Abstract

The present invention relates to a computer implemerited method of performing signal analysis and/or feedback rendering of at least one ECG recording from an ECG recording device, the method comprising steps of : - obtaining a set of reference signals with known and/or assumed properties, determine set limits, such as including a range of values relating to a reference value, such as a percentile, average or median, of the set at a heart beat timing, of the set of reference signals, - receiving at least one ECG heart beat sample from a subject from the ECG recording device, - performing a comparison between and/or determining difference values of the heart beat sample and the set limits, - analyzing the sample comparison and/or the difference values, such as with respect to a value thereof and/or with respect to a disease pointer.

Description

Computer implemented method and system for ECG analysis The present invention relates to a computer imple- mented method of performing signal analysis and/or feedback rendering of at least one ECG recording from an ECG record- ing device. ECG signals have recurring specific properties. Such recurring specific properties comprise the P wave, the QRS complex and the T-wave. Such signals differ from person to person and are dependent on circumstances when recording ECG signals. For the ECG, the placement of the ECG elec- trode on the body is of relevance, but also built of the body, gender, age, weight etc. This leads to different wave shapes per individual, which makes it difficult to quanti- tatively determine whether such repeating signals fall within a certain category, such as the category of a normal ECG recording. ECGs have become increasingly practical for determin- ing deviation of a normal heart function. However, it is also very difficult to interpret ECG signals and a lot of situations exist in which it is very important to make a quick determination as to the requirements a subject might have for fast intervention. In case for example an ambu- lance is faced with a person that might be experiencing a heart condition, it might be partly important to immediate- ly determine to what type of hospital this person is to be brought. It happens a lot that a person that requires imme- diate PCI type care is transported to a hospital that is not equipped for such care. It is a goal according to the present invention to provide a method and system to resolve such mishap. Also, the reliability of interpretation of ECGs is a problem with present ECG systems. To this end, it is a goal according to the present invention to provide a method and system to provide high quality feedback as to assist as- sessment of a specific heart condition of a subject .

As is shown in Fig . 3 , with a limited example set of seven healthy persons , a rather noticeable variation be- tween even these normal ECGs exists also when it pertains to healthy persons . Notwithstanding the variation in wave shapes , the QRS and the T-wave are clearly discernible . It is however a difficulty to determine which of these ECGs represent a normal situation and which would represent the situation with an underlying deviation or disease .

In order to provide a solution to this problem, the present invention provides a computer implemented method of performing signal analysis and/or feedback rendering of at least one ECG recording from an ECG recording device , the method comprising steps of :

- obtaining a set of reference signals with known and/or assumed properties ,

- determine set limits , such as including a range of values relating to a reference value , such as a percentile , average or median, of the set at a heart beat timing, of the set of reference signals ,

- receiving at least one ECG heart beat sample from a subject from the ECG recording device ,

- performing a comparison between and/or determining difference values of the heart beat sample and the set lim- its ,

- analyzing the sample comparison and/or the differ- ence values , such as with respect to a value thereof and/or with respect to a disease pointer .

It is an advantage of such a computer implemented in- vention that data of a single ECG recording, that in itself is hard to determine as to normalcy or deviations from nor- malcy provides an immediate response comprising exactly such indication of normalcy or deviations from normalcy . The set of reference signals preferably comprises at least a large enough number of ECG recordings from a suita- ble control population, preferably healthy persons such that a method according to the present invention provides a result based on suf ficient data to validate the results . Such validation falls in the realm of statistical analysis per se . Specific advantages of specific preferred embodiments according to the present invention are disclosed now by way of example . The present invention relies on the reali zation that a variation of specific categories of generally repet- itive ECG signals is desirable . The method is to be per- formed on a representative set of such signals . Exemplary, the set is formed with ECG recording data from a database ( PTP-Excel ) comprising of for 6500 ECG recordings of per- sons of which also gender and age was known . These ECGs have been divided into 6 age categories for men and women .

According to a preferred embodiment the receiving step comprises steps of data manipulation for normalization of the ECG heartbeat sample .

According to a further preferred embodiment , the re- ceiving step comprises steps of data manipulation for data sampling adj ustment , such as per time unit , preferably to obtain an equal number of samples per time unit between the set of reference signals and the at least one ECG heartbeat s amp 1 e .

According to a further preferred embodiment , method comprises steps for providing an indication of a require- ment of PCI treatment based on the at least one ECG heart- beat sample . Such embodiment is particularly advantageous to provide such indication in case of ambulance emergency help in which transportation to a suitable healthcare fa- cility, such as a hospital , is urgent in case of such indi- cation, and is not particularly urgent in case of not such an indication .

According to a further preferred embodiment , the method comprises steps of performing of a normalization of the ECG heartbeat sample relative to at least 1 of the set of reference signals , preferably a normalization of the at least one ECG heartbeat sample relative to a normalization of the set of reference signals . Such that provides the ad- vantage of being able to directly compare ECG signals based on timings of segments thereof or based on comparable X ax- is values thereof .

According to a further preferred embodiment , the method comprises steps of rendering a feedback signal to provide feedback, such as a visual feedback to a user of a system performing the ECG . With such feedback signal , ver- sion can be provided with interpretations of measurement data that the person would not otherwise determine , or without use of specific analysis work by the person .

According to a further preferred embodiment , the steps of rendering of the feedback signal comprise steps of rendering of signal data of the set of reference signals and/or rendering of signal data of the at least one ECG heartbeat sample . With this , it is provided that the ren- dering of the signal data itself is augmented as for exam- ple a segment thereof in which it exceeds values or com- prises differencial values that are analyzed to be indicat- ed .

According to a further preferred embodiment , the steps of rendering of the feedback signal comprise steps of rendering indicative elements relative to a rendering of signal that of the set of reference signals and/or a ren- dering of signal data of the at least one ECG heartbeat sample . With this , it is provided that either a rendering of the signal data is highlighted by rendering of further indicative signs , such as highlighting or attention provid- ing elements .

According to a further preferred embodiment , the at least one ECG heartbeat sample from a subject from the ECG recording device has been recorded with a predetermined specific electrode placement pattern . It is preferred that there also ECG recording of the set of reference signals was previously recorded with search specific electrode placement . This improves comparability of the recorded ECG data , and does improve the reliability of analysis and in- dications provided by a system according to the invention or embodiments thereof .

According to a further preferred embodiment , the method comprises steps of determining a distance to the me- dian of the set of reference signals of the at least one ECG heartbeat sample . With such determined distance , a di- vergence that is to be noted of the at least one ECG heart- beat sample relative to the set may be a preferred result of the method .

According to a further preferred embodiment , the dis- tance is determined for at least 1 segment of the at least one ECG heartbeat sample , such as of at least one segment between respective fiducials thereof or a part of such seg- ment thereof . Analysis of a segment of the at least one ECG heartbeat sample preferably provides detailed insights , such as information as to a deviation of the at least one ECG heartbeat sample relative to the set . Preferably, such examples directed at changes in direction of the ECG that may be detected based on a subset of ECG leads and signals thereof .

According to a further preferred embodiment , the method comprises steps of determining a reference wave shape distribution based on the set of reference signals . An advantage thereof is that it may be determined whether values of the at least one ECG heartbeat sample deviates from the set or a subset of the set , such as from the medi- an or a range of percentiles .

According to a further preferred embodiment , the method comprises steps of determining whether at least a part of the wave shape of the at least one ECG heartbeat sample has at least one value outside the reference wave shape distribution . With this , specific segments of the wave shape of the at least one ECG heartbeat may be com- pared to respective specific values or subsets of values of the set .

According to a further preferred embodiment , the method comprises steps of determining differential values or a differential graph based on differences between the set of reference signals and the at least one ECG heartbeat sample . Since differential graph provides a highly accessi- ble or easily determinable indication of differences to be determined as any difference is shown as a deviation from the flat line of the x-axis of such graph .

According to a further preferred embodiments , the method comprises steps of assembling a set of reference signals based on past ECG recordings of a single person . With such, the development of the health of a heart of such person may be determined based on its history . With this , the person is compared to himself or herself and when devi- ation starts to occur this is readily indicated .

According to a further preferred embodiment , the at least one ECG heartbeat sample is of the person of the set of reference signals .

A further aspect of the present invention relates to a system for performing signal analysis and/or feedback rendering of at least one ECG recording from an ECG record- ing device , the system comprising :

- at least one processing device with access to at least one memory and/or data store ,

- receiving means for selecting and/or receiving a set of reference signals with known and/or assumed proper- ties ,

- program means configured to function on the at least one processing device to determine set limits , such as including a range of values relating to a reference val- ue , such as a percentile , average or median, of the set at a heart beat timing, of the set of reference signals ,

- receiving means for receiving at least one ECG heart beat sample from a subject from the ECG recording de- vice ,

- program means configured to function on the at least one processing device to perform a comparison between and/or determining difference values of the heart beat sam- ple and the set limits ,

- program means configured to function on the at least one processing device to analyze the sample compari- son and/or the difference values , such as with respect to a value thereof and/or with respect to a disease pointer .

Such aspect according to the present invention has features and advantages that have been described in the above in relation to the above aspect .

A first preferred embodiment of this aspect according to the present invention relates to such system comprising an ECG recording device . In such preferred embodiment , the ECG device may have the rest of the system integrated for ease of operating thereof , such as in an ambulance , ambu- lant facility or hospital .

A further preferred embodiment of this aspect com- prises a user interface, such as a graphic user interface, configured to output a rendering of a feedback based on the comparison or difference values. A further preferred embod- iment of this aspect comprises processing means for per- forming any of the steps according to the aspect as de- scribed in the above .

Further advantages, features and details of the pre- sent invention will be further elucidated on the basis of a description of one or more preferred embodiments with ref- erence to the accompanying figures. Similar yet not neces- sarily identical parts of different preferred embodiments may be indicated with the same reference numerals.

Fig. 1 shows a system overview of an exemplary system according to an embodiment of the present invention.

Fig. 2 A shows an example of a regular ECG recording that is recorded with a 12 lead ECG recording device. Fig.

2 B shows an example of part of an ECG with 2 ways of nor- malizing this part.

Fig. 3 shows an example of 7 regular health ECGs rep- resenting a set of reference signals.

Fig. 4 shows an example of a large set of ECGs repre- senting a set of reference signals of a P wave segment for male and female subjects.

Fig. 5 shows an example of a large set of ECGs repre- senting a set of reference signals of QRS and T-wave seg- ments for male and female subjects.

Fig. 6 shows an example of a large set combining Fig.

4 and Fig. 5, generally showing a lack of amplitudes above

3 mV in 99.9% of ECGs.

Fig. 7 shows a graph of a large set of ECGs repre- senting a set of reference signals with a subject to be compared according to the present invention showing a devi- ation . Fig . 8 shows a graph with an example of normalization for application with the present invention limited to a P wave segment .

Fig . 9 shows a graph of a large set of ECGs repre- senting a set of reference signals recorded with the fee to lead of the ECG device , generally indicating set limits ex- amples , such as the 1^st percentile , the median and the 99^th percentile .

Fig . 10 shows a graph with supplements and the values of a heartbeat sample to be assessed according to the pre- sent invention, as well as a graphic difference values .

Fig . 11 shows a distance to median graph of a sample relative to a set of reference signals .

Fig . 12 shows

A first preferred embodiment ( Fig . 1 ) according to the present invention relates to a system for performing ECG analysis , preferably according to a method according to the present invention . The system comprises an optional ECG recording device 6 that is pre ferably provided with a pre- ferred 12 leads configuration to connect to a subject as of which at least one ECG heartbeat sample is taken by means of the ECG recording device 6 . A computer 5 comprises at least one processor and at least one memory and at least one data store for long term storage . This computer is preferably connected to a data store for retrieving a set of reference signals , such as of previous relevant ECG ses- sions of patients from a peer group or from the same pa- tient in the past , based on selection criteria to be deter- mined by the user of the computer or that is generally based on decision logic based on the type of subject that is present . An example of a representative set preferably is based on ECGs of normally healthy persons , such as ob- tainable from the PTB-XL database comprising ECGs of which also gender and age is known . The computer comprises a graphical user interface such as by means of a display 7 and can be operated by means of input means such as a key- board 8 and the mouse 9 . With this , a user can make the computer 5 perform a method according to the present inven- tion and operate the ECG device to obtain a current ECG re- cording of a person that is to be assessed based on the set of reference signals while using the computer implemented method according to the present invention .

Parameters that are relevant to assemble the set of reference signals include age , health, weight , gender, a known heart condition etc . For each of such categories , a variation map per ECG signal can be formed or rendered . The wave shapes of such ECG, although periodic, also depends on the heartbeat . The higher the heartbeat , the faster the heart is to function and the shorter the heartbeat becomes , such as the time between the P wave and data QRS complex, as well as the duration of the T-wave and the QRS complex albeit to a lesser extent . With this , all ECGs have a dif- ferent length . It is preferred to form the set of reference signals based on normali zed ECG signals , such as by equal- izing the length of the ECG signals per reading, heartbeat or a segment thereof , or by making the number of samples the same for each of those . For P wave , it is preferred, as it is caused by the electric activation of the atria and the QRS-T-wave by deactivation and replarization of the chambers , to create two reference wave shape graphs , one for the P wave and one for the chamber activity, such as the QRS-T-wave .

It is preferred that the ECGs are split up in seg- ments , such as for the P-wave or QRS-T-wave . Thi s means that the beginning and the ending of the P wave and begin- ning and ending of the QRS and ending of the T-wave is de- termined . This may be performed automatically during pro- cessing of the present invention, but it is also possible to rely on such earlier determinations that have been stored previously, such as with the set of reference sig- nals .

Furthermore , all ECGs are sampled in a way that each respective set comprises the same number of samples , such as 200 samples for the P wave from the beginning of the P wave to the beginning of the QRS and 400 samples for the normal QRS-T-wave part of respective ECG readings . An exam- ple thereof is that when a QRS-T-wave , that is 420 ms long, will be compressed to the equivalent of 400 ms per sample . As the most variable part of the QRS-T-wave is in the T- wave , it is preferred that most samples are made in the T- wave part until the maximum number of samples is reached, such as for example these 400 samples . In the example of Fig . 2 , the shortest wave shape 21 with a QRS-T duration of 320 ms represents the original ECG recording . The middle wave shape 22 has been extended to the 400 ms shape by nor- mali zing only the T-wave . For this , the first 110 ms of the QRS-wave have been left the same . The most right shape 23 has been normali zed by extending the whole QRS-T shape to the 400 ms shape . In such way, all QRS-T shapes may be nor- mali zed to perform the same number of samplings . Fig . 8 shows the process of normalization and adapting of the graphs as indicated in a more detailed way .

The value of normalizing up to 400 is an arbitrary one with understandability within the field .

Any alternative resampling may be performed, such as by also providing the interval between the end of the QRS and the peak of the T-wave a fixed length and providing the remaining part of the T-wave with a di f ferent fixed length .

A probability profile may be created with the pre- ferred plotting of all ECGs from a certain category in a single graph, such as shown in respective Figs . 4 , 5 , 6 , 7 , 9 , the probability profile indicating the occurrence of a certain amplitude at a certain time of the ECG recordings . With this , preferred inter-quartile or inter-percentile ranges may be determined . Lines that are reconstructed in such way, indicate how large the chance is that a certain amplitude occurs in the QRS , the T-wave , P wave or the PR interval .

Fig . 4 shows an average P wave of a large group of aligned or normali zed ECGs . The top shows males between 18 and 90 years old and the bottom shows females between 18 and 90 years old . It is preferred that outliers are re- moved . A normal ECG preferably lies within a range of 1 - 99% , but such ranges may be determined based on the purpose of the application of the invention . Fig . 5 shows a QRS and T-wave of a large number of aligned ECG recordings . The top shows males between 18 and 90 . The bottom shows females be- tween 18 and 90 . The QRS amplitudes in males may be higher than in females . In males , amplitudes above 3 mV are less than .25% of the ECGs . Figures 6 shows a combination of the preceding male and female graphs .

Fig . 7 shows an example of a feedback that might be provided by the system applying the method according to the present invention or embodiments thereof . Within a set of re ference signals , the signal of a subject is drawn . What is highly noticeable , is that in the graph of the VI lead of the ECG, the person exhibits levels well below the medi- an of the population, whereas for the V2 lead, the person exhibits levels well above the median of the population . From this , the comparison between and/or determining dif- ference values of the heart beat sample and the set limits provides a clear and immediate indication that this person might be experiencing a condition that makes him deviate so clearly between 2 leads of the same ECG.

Fig. 10 shows the ECG of a subject, generally the middle line 103, together with the set limits 101,102 of the set of reference signals. From this, difference values are determined which means that if the values of the ECG of the subject are above the upper limit 101 and if the values of the ECG of subject are below the lower limit 102, the difference 104 between the respective upper limit and lower limit will provide a value.

If the values of the ECG of the subject are above the upper limit and the values of the upper limit are subtract- ed therefrom, a positive graph will be shown in the differ- ence value graphs at the bottom half of Fig. 10. This is generally the case in the leads AVR, VI, V4 and II, III aVF.

After values of the ECG of the subject are below the lower limit and values of the subject are subtracted from the lower limit, a negative graph will be shown in the dif- ference value graphs at the bottom half of Fig. 10. This is relatively more clearly the case in V4, II, V2, V5. avF and V3.

These graphs make it very clear that the values of this person and these locations deviate outside of the first and 99^th percentile of the set of reference signals. When this happens, these graphs are shown and a warning is rendered. Such warning maybe embodied as the shown devia- tion of line 104 from the horizontal 0 axis, alternatively, a further indication in the graphical user interface is considered, as well as a printed warning or an audible warning .

Fig. 10 also shows a further representation that may be rendered in order to show a deviation of a present ECG 103 relative to other ECGs of a set . If the ECG comprises segments 103 ' that deviate from the set , this is indicated by a superimposed rectangular shaped indication 105 or block wave . This shows the indication of a deviating ECG relative to the set in a very clearly discernible manner . With a system of comparison according to the invention or an embodiment , all new ECGs or ECG measurements , such as comprising periodic signals with comparable properties , may be quickly classi fied . Depending on the used data to con- struct the reference wave shape distribution, distinct classi fications may be provided .

Fig . 11 shows a graph of a distance to median of the present ECG relative to the set of reference signals . It is preferably interpreted that the distribution lines may be used to determine in which so-called quartile the ECG sam- ple is located . This may be inferred from having signals ranging between minus 0 . 5- 1 . 5 , wherein zero indicates that the presently recorded ECG is within the median class and . 5 indicates it is outside the distribution . Also in this graph, a graph well above or below zero provides an indica- tion, such as the graphical feedback, that a deviation from the median is being perceived .

Fig . 12 shows differences with the median of the set of reference signals . A patient that has an intermittent di f ferent right bundle branch block, possibly with a septal activation, which is the case for heartbeat one and without septal activation, which is the case for heartbeat 2 . These are successive heartbeats that clearly indicate the rele- vance of the assessment with this graph . Without the septal activation, the ST segment along several ECG leads is high, whereas when there is septal activation, dips in this ele- vated ST segment are visible . This clearly indicates how granular the present invention has its ef fect in providing visible representations of heart conditions.

Fig. 13 indicates large differences in QRS amplitude in the ECG, but also in the STT segment. Also large differ- ences are shown in CineECG positions in the STT segment.

Deviation percentages per segment are exemplary as follows :

Cine Ecg Normal Outside

QRS : 82% 18%

ST : 0% 100%

T-wave : 0% 100%

ECG amoplitudes

QRS : 14% 86%

ST : 0% 100%

T-wave : 0% 100% with

Fig. 14 generally indicates large differences in the

ST segment just after the QRS. Also significant differences in Cine ECG positions in the STT segment are shown.

Deviation percentages per segment are exemplary as follows :

Cine Ecg Normal Outside

QRS : 100% 0%

ST : 11% 89%

T-wave : 0% 100%

ECG amoplitudes

QRS : 56% 44%

ST : 0% 100%

T-wave : 37% 63%

Figures 15 and 16 provide clearly visible measure- ments and explanatory tables of analysis .

Fig. 15 relates to a case of ARVC. ARVC is a heredi- tary heart disease in which there is an increased risk of arrhythmias (arrhythmogenic) due to fatty and scarring of the right and to a lesser extent left ventricle (ventricu- lar cardiomyopathy) . Analysis data comprises:

QRSD: 90 ms

Tpeak: 369 ms

QT: 457 ms

QTc: 448 ms

RR int: 1040 ms

Cine Ecg Normal Outside

QRS: 100 % 0 %

ST: 100 % 0 %

T-wave: 78 % 22 %

ECG amoplitudes

QRS: 100% 0%

ST: 100% 0%

T-wave: 66% 34%

HB system: normal discordant QRST: 0%, 155

ST deviation: 0.14 mV relative ST: 0.094 segment: 8, 8 experiment: 50%, 3 angle early ST-TpTe: 129 deg

Fig. 16 relates to a case of BrS 11 Type 1 baseline and after ablation. Analysis data for Fig. 16A comprises: QRSD: 121 ms

Tpeak: 301 ms

QT: 452 ms

QTc: 539 ms

RR int: 702 ms

Cine Ecg Normal Outside

QRS: 59 % 41 %

ST: 100 % 0% T-wave: 100% 0%

ECG amoplitudes QRS: 48% 52%

ST: 70% 30%

T-wave: 100% 0%

HB system: other discordant QRST: 0%, 129 ST deviation: 0.95 mV relative ST: 0.42 lead ST 60-80 J-point

VI : 0.26 0.56 mV

V2: 0.27 0.34 mV

V3: 0.16 -0.14 mV segment: 8, 8 experiment: 50%, 3 angle early ST-TpTe: 172 deg

Analysis data for Fig. 16B comprises:

QRSD: 112 ms

Tpeak: 262 ms

QT: 377 ms

QTc: 448 ms

RR int: 708 ms

Cine Ecg Normal Outside

QRS: 74 % 26 %

ST: 100 % 0%

T-wave: 100% 0%

ECG amoplitudes

QRS: 67% 33%

ST: 100% 0%

T-wave: 100% 0%

HP system: IVCD septal discordant QRST: 0%, 81 ST deviation: 0.76 mV relative ST: 0.32 lead ST 60-80 J-point

VI : 0.14 0.02 mV

V2: 0.22 0.07 mV

V3: 0.16 -0.02 mV segment: 8, 8 experiment: 60%, 3 angle early ST-TpTe: 66 degFig. 17-Fig. 20 provide a description of steps that are described below. In Fig. 17, step 100 depicts data collection of a certain type, e.g. normal healthy humans, Brugada patients, people with an is- chemic event in the Left anterior descending etc. Every ECG is preferably be recorded under the same circumstances to make the reference as homogeneous as possible. In step 200, recorded ECG representative ECG beats, which may be atrial or ventricular, depending on the purpose at hand, are se- lected or constructed (median beat) . These ECG signals are used to build the reference waveform distribution. In step 300, the reference waveform distribution is stored with suitable description to a database. 400 represents the da- tabase of reference waveform distributions. In step 500, a recorded ECG is compared to a reference waveform distribu- tion of a specific type selected from the database and de- viating waveform segments are identified. In step 600, based on the deviating waveform segments a diagnostic mes- sage is preferably produced. For instance when the ST seg- ment, just after the end of the QRS, is deviating signifi- cantly acute myocardial infarction (heart attack) might be expected .

In Fig. 18, a description is given of preferred re- constructing of reference waveforms. In step 210, importing of waveform data of a set of reference signals such as of many subjects (can also be CineECG data) is performed. In step 220, a heartbeat is selected, preferably this may be a signal averaged beat, i.e. the average of multiple beats with similar waveform. In step 230, steps are performed for determining beat fiducials: the fiducials of a single ECG beat from the 12 lead CEG are preferably derived from the root mean square (RMS) of all ECG signals measured. The ECG fiducials comprise the QRS onset, QRS offset, the Pwave On- set, P wave Offset, T wave Peak and T wave end. The QRS on- set is preferably defined as the time when subsequent ECG samples have an increasing value for at least 10 ms. The QRS offset is preferably defined as the time when the RMS amplitude is lowest between 80 and 200 ms after the detect- ed QRS onset. QRS90 is preferably defined as the time 90 ms after QRS onset. J-Point 30 is preferably defined as the time 30 ms after the QRS offset, and the Q-point is prefer- ably defined by the intersection point at the time axis and the upslope tangent between the T-peak and the mid- ampli- tude T-wave (oranges lines) . Similarly, the T-wave end is preferably defined by the inter section point at the time axis and the downslope tangent between the T wave peak and the mid T wave amplitude (blue lines) . Fiducial of the P wave are determined based on the low RMS amplitudes. In step 240, resampling is performed of preferably any or every waveform (atrial or ventricular) to the predefined number of samples and store temporarily until no more wave- forms can be added (250) . In step 260, a reference is cre- ated based on preferably waveforms (amplitude) and/or wave- form distribution for the supplied waveform data.

In Fig. 19, an example is provided of an analysis and/or diagnostic module. In step 510, the ECG is obtained or recorded, preferably by means of connecting the wires of the ECG recorder to the electrodes on the chest of the pa- tient. For the standard 12 lead ECG this the left and right arms, the left foot and the precordial leads V1-V6 over the heart as shown below. When a 3D camera is available, the ECG electrodes might be recorded as well to ensure a homo- geneous database for reference building. In step 520, the heartbeat is selected, this pertains preferably to a signal averaged beat, i.e. the average of multiple beats with sim- ilar waveform. In step 530, the fiducials of the heartbeat are preferably determined: the fiducials of a single ECG beat from the 12 lead CEG are automatically derived from the root mean square (RMS) of all ECG signals measured. The QRS onset is preferably defined as the time when subsequent ECG samples have an increasing value for at least 10 ms. The QRS offset is preferably defined as the time when the RMS amplitude is lowest between 80 and 200 ms after the de- tected QRS onset. QRS90 is defined as the time 90 ms after QRS onset. J-Point 30 is defined as the time 30 ms after the QRS offset, and the Q-point is defined by the intersec- tion point at the time axis and the upslope tangent between the T-peak and the mid- amplitude T-wave (oranges lines) . Similarly, the T-wave end is defined by the inter section point at the time axis and the downslope tangent between the T wave peak and the mid T wave amplitude (blue lines) . Fiducial of the P wave are determined based on the low RMS amplitudes. In step 540, the required reference waveform distribution is selected from the database and resample this database data to have the same number of samples as the measured ECG beat (atrial or ventricular) . Generate then a difference map (described above) .

In Fig. 20, a preferred embodiment regarding determi- nation of differentials is shown. In step 610, segments where the newly recorded waveforms are outside of the cho- sen waveform distribution are determined. These are depict- ed in the graphs of Fig . 10 . In step 620 , it is determined whether segments are outside of the limits . In case this is true , a warning is rendered in step 640 . In case this is not true , the data is presented as matching the provided waveform distribution .

Definitions .

A set limit according to the present invention is de- fined as a limit relating to a set of reference signals . Such a set limit is preferably determined as a value of an ECG recording at a specific timing during performing of an ECG recording during a heartbeat . For the purpose of the present invention, a set limit is further preferably de- fined as the highest value of the set at such specific tim- ing .

A set limit preferably is determined based on normal- ized ECG recordings . Normali zed ECG recordings are prefera- bly defined as ECG recordings of which a duration of a heartbeat or a segment thereof is normali zed, such as that segments from several ECG recordings from the said have generally the same length . Preferably this means that each ECG recording of the set as well as a for comparing meas- urement values relative to such normali zed ECG recordings .

A fiducial of an ECG according to the present inven- tion comprise the QRS onset , QRS of fset , the Pwave Onset , P wave Of fset , T wave Peak and T wave end .

Clause

Computer implemented method of performing signal analysis and/or feedback rendering of at least one ECG re- cording from an ECG recording device , the method comprising steps of :

- obtaining a set of reference signals with known and/or assumed properties ,

- determine set limits , such as including a range of values relating to a reference value, such as a percentile, average or median, of the set at a heart beat timing, of the set of reference signals,

- receiving at least one ECG heart beat sample from a subject from the ECG recording device,

- optionally performing a comparison between and/or determining difference values of the heart beat sample and the set limits,

- optionally analyzing the sample comparison and/or the difference values, such as with respect to a value thereof and/or with respect to a disease pointer.

A further example is provided in Fig. 21, in which a subject with chest pain appears suitable for Percutaneous coronary intervention. An initial prior art ECG shows no significant signs of ST elevation. Only minor ST depression in V2, all other ST values were within normal range. The initial ECG is as depicted.

A CineECG as depicted in Fig. 22 shows a clear devi- ating path of an STT segment, directed towards the LV lat- eral wall, an area associated with the Left Circumflex (LCx) , specifically the OM branch. In this vessel a stent was previously placed.

By comparing the CineECG with a normal distribution of CineECG positions, such as depicted in Fig. 23, it is shown clearly that the STT segment 231 is largely out of range. This normal CineECG distribution is suitable to as- sist users of the system to identify abnormal (Cine) ECG waveform patterns, like in in this case of an acute occlu- sion of the OM branch of the LCx

Distance from the median maps, such as shown in Fig. 24, also shows that the ST segment in VI V2, II , III and aVF are elevated whereas the rest of the lead show a near median ST segment. This points towards left lateral direc- tion, and thus confirms the conclusion above .

The distance from the outer distribution lines , such as shown in fig . 25 , shows also that the ST segment of V2 is deviating, but not the other leads . This provides an in- dication that figures 24 and 25 complement each other in identi fying deviating amplitudes in di f ferent leads . Lead V2 is the lead opposite of the left lateral wall an thus should show most ST depression ( negative amplitudes ) .

Fig . 26 provides the results of a method in relation to the history of a single person . In this case , the set pertains to a part of the history of the person in which a cardiac event has not yet occurred . The median value is in- dicated by reference numeral 260 . Values that fall within 85% or more correspondence with a reference ECG measure- ment , such as the 1 st ECG measurements of the set , are in- dicated by the reference numeral 262 . Values that fall within a 60- 85% correspondence with this reference ECG measurement are indicated by the reference numeral 263 . The last ECG measurement is indicated by reference numeral 261 the last . This last ECG measurement , corresponding with the at least one ECG heart beat sample according to the inven- tion or an embodiment thereof falls outside of these ranges of 60- 85% , and 85% or more , and this indicates an ECG meas- urement of which at least a segment is to be indicated as deviating according to the present invention or in embodi- ments thereof . In a graphical representation this is thus indicated as a distinct line that may be represented with a distinct color of the line representing the ECG heart beat sample in the graphical representation relative to the ECG heartbeat samples that are part of the set .

Fig . 27 provides graphical representations of the person after a cardiac arrest has occurred . It is clear that the values of the ECG are relatively upset by the oc- currence of the cardiac arrest previously . It is a specific point of the invention that by acting upon the result of fig . 26 , situations such as fear 27 may be prevented .

Fig . 28 provides data of an improved results relative to fig . 27 after surgery including inflating of a balloon in a heart artery to obviate underlying causes of the car- diac arrest . The representations of the later ECG measure- ments provide a much more quiet situation than before the surgery after the cardiac arrest . This provides an indica- tion of success of the operation .

It is a point of the present invention or preferred embodiments thereof that a patient with a genetic deviation usually does not have typical ECG changes that go with this genetic deviation (phenotype versus genotype ) . The pheno- type does not even always come into being even though the genetic deviation is present . In case the symptoms do start to appear, often times a defibrillator is implanted . Rather or when someone that ful fills the criteria is hard to pre- dict , but a point of the present invention or embodiments thereof is that ECG comparisons throughout the time pro- vides an obj ective measure to detect subtle changes in the ECG or segments thereof . It is a point of the present in- vention that such changes in segments of the ECG may be de- tected by waveform analysis . If a distribution is deter- mined of a waveform of the baseline ECGs , then it is an in- side of the present invention that deviations of the wave- form of the ECG may be determined based on the waveforms itself . When deviations become " large" enough, the present invention or embodiments thereof provides an indication to perform a more thorough examination of search subject, such as by means of a higher frequency of subsequent instances of taking measurements towards at least one ECG heartbeat sample the . In other words , when deviations occur in ECGs when taken for instance annually or biannually, a change of regime towards 4 times or 12 times a year might be a 1^st step towards further treatment .

Fig. 29 provides a flow diagram of a method specifi- cally intended to perform a comparison of the at least one ECG heart beat sample with a set of reference signals based on previous ECG heartbeat samples of the same person. The set is imported in step 700' . In step 610' , it is deter- mined whether parts of signal waveforms are outside of val- ues of the set or a waveform distribution thereof. In step 620' a percentile correlation is determined between the latest or the at least one ECG heartbeat sample and the set. An example of calculations that are performed for search steps is making use of Pearson's correlation coeffi- cient. With this, X = the reference (Cine) ECG signal, Y = the other (Cine) ECG signal with the same number of sam- ples, wherein preferably use is made of either all ECG sig- nals or the 3 CineECG signals, being apex ->base, right- >left, posterior->anterior ) .

Pearson's correlation coefficient is used as follows. The sample is preferably represented by r_xy and can be re- ferred to as the sample correlation coefficient authority sample Pearson correlation coefficient. The following for- mula is used for substituting estimates of covariances and variances based on the sample in the formula. Use is made of paired data in the following format.

{ (x1,y1) (x_n,y_n) } forming the n pairs, r_xy for the following formula:

where

The present invention is described in the foregoing on the basis of preferred embodiments . Different aspects of different embodiments are expressly considered disclosed in combination with each other and in all combinations that on the basis of this document , when read by a skilled person of the area of skill , fall within the scope of the inven- tion or are deemed to be read with the disclosure of this document . These preferred embodiments are not limitative for the scope of protection of this document . The rights sought are defined in the appended claims .

Claims

1 . Computer implemented method of performing signal analysis and/or feedback rendering of at least one ECG re- cording from an ECG recording device , the method comprising steps of :

- obtaining a set of reference signals with known and/or assumed properties ,

- analyzing the sample comparison and/or the di f fer- ence values , such as with respect to a value thereof and/or with respect to a disease pointer .

2 . Method according to claim one in which the receiv- ing step comprises steps of data manipulation for normali- zation of the ECG heartbeat sample .

3 . Method according to claim one wherein the receiv- ing step comprises steps of data manipulation for data sam- pling adj ustment , such as per time unit , preferably to ob- tain an equal number of samples per time unit between the set of reference signals and the at least one ECG heartbeat sample .

4 . Method according to any of the preceding claims wherein the steps of performing a sample comparison com- prise steps for providing an indication of a requirement of PCI treatment based on the at least one ECG heartbeat sam- ple .

5 . Method according to claim 4 wherein the method comprises steps of performing of a normalization of the ECG heartbeat sample relative to at least 1 of the set of ref- erence signals , preferably a normalization of the ECG heartbeat sample relative to a normalization of the set of reference signals .

6 . Method according to any of the preceding claims comprising steps of rendering a feedback signal to provide a feedback, such as a visual feedback to a user of a system performing the ECG .

7 . Method according to claim 6 wherein the steps of rendering of the feedback signal comprise steps of render- ing of signal data of the set of reference signals and/or a rendering of signal data of the at least one ECG heartbeat s amp 1 e .

8 . Method according to claim 6 or 7 wherein the steps of rendering of the feedback signal comprise steps of ren- dering indicative elements relative to a rendering of sig- nal that of the set of reference signals and/or a rendering of signal data of the at least one ECG heartbeat sample .

9 . Method according to one or more of the preceding claims wherein the at least one ECG heartbeat sample from a subject from the ECG recording device has been recorded with a predetermined specific electrode placement pattern .

10 . Method according to one or more of the preceding claims comprising steps of determining a distance to the median of the set of reference signals of the at least one ECG heartbeat sample .

11 . Method according to claim 10 wherein the distance is determined for at least 1 segment of the at least one ECG heartbeat sample , such as of at least one segment be- tween respective fiducials thereof or a part of such seg- ment thereof .

12 . Method according to any of the preceding claims comprising steps of determining a reference wave shape dis- tribution based on the set of reference signals .

13 . Method according to claim 12 comprising steps of determining whether at least a part of the wave shape of the at least one ECG heartbeat sample has at least one val- ue outside the reference wave shape distribution .

14 . Method according to any of the preceding claims comprising steps of determining differential values or a differential graph based on differences between the set of reference signals and the at least one ECG heartbeat sam- ple .

15 . Method according to any of the preceding claims comprising steps of assembling a set of reference signals based on past ECG recordings of a single person .

16 . Method according to claim 15 in which the at least one ECG heartbeat sample is of the person of the set of reference signals .

17 . System for performing signal analysis and/or feedback rendering of at least one ECG recording from an ECG recording device , the system comprising :

18 . System according to the preceding claim compris- ing an ECG recording device .

19 . System according to the preceding claims 17 or 18 comprising a user interface , such as a graphic user inter- face , configured to output a rendering of a feedback based on the comparison or difference values .

20. System according to the preceding claims 17, 18 or 19, comprising processing means for performing any of the steps according to any of the claims 1-16