Computer-implemented process and related system for electrocardiogram signal analysis
The computer-implemented electrocardiogram signal analysis process segments and identifies key features like QRS complexes and T waves using derivative calculations, addressing the limitations of existing methods to enhance accuracy and reliability.
Patent Information
- Application Number
- PCT/IB2026/051989
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-03-03
- Filing Date
- 2026-03-02
- Publication Date
- 2026-11-12
AI Technical Summary
Existing electrocardiogram signal analysis methods are limited in their ability to automatically and precisely identify the morphology of waves and complexes, and are not capable of simultaneously identifying all components of the electrocardiogram trace, leading to reduced accuracy and reliability.
A computer-implemented process that segments electrocardiogram signals into cardiac cycles, identifies key features like QRS complexes, T waves, and J points through derivative calculations and inflection points, and aligns signals with an isoelectric line to accurately determine morphological characteristics.
Enables precise and automated identification of electrocardiogram signal components, improving analysis accuracy and reliability by accurately determining J points and other morphological features.
Smart Images

Figure IB2026051989_12112026_PF_FP_ABST
Abstract
Description
[0001] COMPUTER-IMPLEMENTED PROCESS AND RELATED SYSTEM FOR ELECTROCARDIOGRAM SIGNAL
[0002] ANALYSIS
[0003] DESCRIPTION
[0004] FIELD OF THE INVENTION
[0005] The present invention relates to a computer-implemented electrocardiogram signal analysis process and a system capable of analysing electrocardiogram signals by carrying out said analysis process. For example, the invention may have advantageous application in the medical sector.
[0006] STATE OF THE ART
[0007] Equipment for acquiring and analysing electrocardiogram signals and configured to monitor and record the electrical activity of the heart through electrodes is known in the sector of cardiac monitoring. The acquired signal is then filtered to eliminate noise, amplified and digitised to enable its analysis. The digitised signal is subsequently analysed to identify P and T waves and ST segments, as well as the QRS complex of a heartbeat, enabling physicians to diagnose basic heart conditions such as arrhythmias and infarctions.
[0008] In a first example described in US7477936B2, an electrocardiogram signal analysis process is envisaged which provides for the acquisition and filtering of signal components, followed by the identification of components for determining parameters such as symmetry, ratio and kurtosis of the signal to discriminate between individuals with Long QT HERG syndrome and normal ones.
[0009] In a second example described in WO 2020 / 159738 A1, there is envisaged a system for determining the J point of an electrocardiogram trace by identifying the S and T peaks and a reference line. The system is configured then to locate the J point by geometric analysis of the amplitude ratios between the S and T points and the reference line.
[0010] In a third example, document US2011040200A discloses a process for analysing an ST segment of an electrocardiogram signal to identify cardiac anomalies such as ischaemia and myocardial infarction, including signal filtering and identification of R, T, J and P points.
[0011] In a fourth example described by US 20050027202 A1, there is envisaged a process for electrocardiogram signal analysis, which provides for calculating an index value based on the standard deviation and mean value to identify possible cardiac dysfunctions. The process further provides for the identification of ECG signal components such as Q, S, and J points and R peaks to determine cardiac anomalies.
[0012] Although the above-described solutions enable an analysis of electrocardiogram signals, the Applicant has found that such solutions show limitations and drawbacks. In particular, the known processes of electrocardiogram signal analysis are not capable of automatically and precisely identifying the morphology of waves and complexes of electrocardiogram signals. Furthermore, such solutions are applicable only with certain configurations of the electrocardiogram signal and are not capable of simultaneously identifying all the waves and complexes of the trace, thus limiting the accuracy and reliability of the analysis.
[0013] SCOPE OF THE INVENTION
[0014] The aim of the present invention is thus to resolve at least one of the drawbacks and / or limitations of the prior solutions. A first object of the present invention is to provide an extremely accurate system and process for automatically identifying significant points of the electrocardiogram signal. In particular, it is an aim of the present invention to provide a system and a process capable of determining with precision and classifying the morphology of an electrocardiogram signal, in particular the J point.
[0015] It is a further aim of the present invention to provide a system and a method capable of effectively filtering out noise from the sampled electrocardiogram signals.
[0016] These aims and still others, which will become more apparent from the following description, are substantially achieved by a computer-implemented process and an associated system of electrocardiogram signal analysis in accordance with one or more of the appended claims and / or of the following aspects.
[0017] SUMMARY
[0018] In an aspect 1, there is provided a computer-implemented process for analysing at least one electrocardiogram trace, optionally relating to one or more leads of an electrocardiogram of a patient's heart, said process comprising the steps of:
[0019] - acquiring at least one electrocardiogram signal which is variable as a function of time, said electrocardiogram signal being a sampled digital signal represented by a series of discrete points,
[0020] - segmenting the electrocardiogram signal into one or more cardiac cycles, each of which representative of a single cardiac cycle,
[0021] - identifying, for each of said one or more cardiac cycles, a QRS complex (QRS) as a segment of the electrocardiogram signal representative of ventricular depolarisation,
[0022] - determining, for each of said one or more cardiac cycles, a T wave (T-wave) as a segment of the electrocardiogram signal comprised between the QRS complex (QRS) and a terminal segment of a respective one of said one or more cardiac cycles, said T wave being representative of ventricular repolarisation.
[0023] In an aspect 2 according to the preceding aspect, the process comprises a step of determining a J point (J-point) of said series of discrete points of the electrocardiogram trace comprised between the QRS complex and the T wave, wherein a change of concavity occurs in the same electrocardiogram trace. In an aspect 3 according to either of the preceding aspects, the step of determining the J point (J-point) comprises the sub-steps of:
[0024] - determining, among said series of discrete points, an S point (S-peak) of the QRS complex, representative of depolarisation of basal portions of the patient's heart,
[0025] - determining, among said series of discrete points, a T point (T-peak) of the T wave (T-wave) representative of ventricular repolarisation of the patient's heart.
[0026] In an aspect 4 according to the preceding aspect, the step of determining the J point (J-point) further comprises the sub-steps of:
[0027] - determining a J point search segment of the electrocardiogram trace comprised between the S point (S-peak) and the T point (T-peak),
[0028] - calculating, among said series of discrete points, one or more inflection points of the J point search segment,
[0029] - identifying the J point (J-point) in one of said one or more inflection points.
[0030] In an aspect 5 according to the preceding aspect, calculating said one or more inflection points of the J point search segment comprises:
[0031] - calculating a second derivative of the J point search segment of the electrocardiogram signal,
[0032] - calculating one or more noteworthy J points, wherein the second derivative of the J point search segment is zero, does not exist or is not defined,
[0033] - identifying, between said one or more noteworthy J points, an inflection point if a change of sign of the second derivative of the J point search segment occurs in said one or more noteworthy J points.
[0034] In an aspect 6 according to any one of the preceding aspects, the process comprises a step of determining the presence of an R' peak (R'-peak) representative of a second ventricular depolarisation, said R' peak being located immediately following the S point (S-peak) and in the QRS complex.
[0035] In an aspect 7 according to the preceding aspect, the electrocardiogram signal has a first derivative equal to zero at the R' peak (R'-peak) and has, in a point immediately following the R' peak (R'-peak), a negative first derivative. In an aspect 8 according to either of the two preceding aspects, if the step of determining the R' peak (R'-peak) gives a positive outcome, the J point search segment of the electrocardiogram trace is comprised between the R' peak and the T point (T-peak).
[0036] In an aspect 9 according to any one of the five preceding aspects, if the step of calculating one or more inflection points of the J point search segment results in the identification of a single inflection point, then the step of identifying the J point envisages marking said single inflection point as the J point (J-point). In an aspect 10 according to any one of the preceding aspects 4 to 8, if the step of calculating one or more inflection points of the J point search segment results in the identification of a plurality of inflection points, then the step of identifying the J point envisages marking as the J point (J-point):
[0037] - the point with the first derivative of the J point search segment equal to zero, or
[0038] - the point, among said discrete points, with a maximum curvature comprised between the R' peak or the S point, and the T point (T-peak).
[0039] In an aspect 10 according to any one of the preceding aspects, the T wave (T-wave) is comprised between an initial point of the T wave (T-onset) of said series of discrete points and a final point of the T wave (T-endset) of said series of discrete points.
[0040] In an aspect 12 according to any one of the preceding aspects, the step of determining the T wave (T-wave) comprises marking as the final point of the T wave (T-endset), a stationary T point of said series of discrete points, wherein a first derivative of the segment of the electrocardiogram signal comprised between the T point (T-peak) and a terminal segment of a respective one of said one or more cardiac cycles is zero.
[0041] In an aspect 13 according to any one of the preceding aspects, determining the T wave (T-wave) comprises marking as the initial point of the T wave (T-onset) a point of said series of discrete points defining a JT segment (JT-wave) of the electrocardiogram trace comprised between the J point (J-point) and the T point (T-peak), wherein a change ofconcavity occurs in the same JT segment (JT-wave). In an aspect 14 according to the preceding aspect, the step of marking the initial point of the T wave (T-onset) comprises:
[0042] - calculating a second derivative of the JT segment (JT-wave),
[0043] - analysing the sign of the second derivative of the JT segment (JT-wave) at the T point (T-peak) in order to determine a concavity of the cardiac cycle at the T point.
[0044] In an aspect 15 according to the preceding aspect, marking the initial point of the T wave (T-onset) comprises further calculating one or more critical T points of said series of discrete points, wherein the second derivative of the JT segment (JT-wave) is equal to zero, does not exist or is not defined. In an aspect 16 according to the preceding aspect, marking the initial point of the T wave (T-onset) comprises analysing the sign of the second derivative of the JT segment (JT-wave) at each of said one or more critical T points to determine a concavity of the cardiac cycle in said one or more critical T points. In an aspect 17 according to the preceding aspect, marking the initial point of the T wave (T-onset) comprises identifying, between said one or more critical T points, the initial point of the T wave (T-onset) as the critical T point where the curvature of the JT segment (JT-wave) has a maximum absolute value and concavity of sign opposite the one calculated at the T point (T-peak). In an aspect 18 according to any one of the preceding aspects, determining the T-point envisages marking the T point (T-peak) as an absolute maximum or minimum point of the T wave (T-wave). In an aspect 19 according to any one of the preceding aspects determining the T point (T-peak) comprises the substeps of calculating, among said series of discrete points, one or more stationary T points wherein a first derivative of the segment of the electrocardiogram signal comprised between the QRS complex (QRS) and a terminal segment of a respective one of said one or more cardiac cycles is zero. In an aspect 20 according to the preceding aspect, if the step of calculating said one or more stationary T points results in the identification of a single stationary point, then the process envisages marking said single stationary point as the T point (T-peak).
[0045] In an aspect 21 according to aspect 19, if the step of calculating said one or more stationary T points results in the identification of a plurality of stationary points, then the process comprises determining a stationary T point with the maximum value and a stationary T point with the minimum value among the plurality of stationary T points. In an aspect 22 according to the preceding aspect, if the step of calculating said one or more stationary T points results in the identification of a plurality of stationary points, then the process further comprises a step of marking as the T point (T-peak) the stationary T point with the minimum value if the stationary T point with the maximum value is negative relative to an isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials. In an aspect 23 according to either of the two preceding aspects, if the step of calculating said one or more stationary T points results in the identification of a plurality of stationary points, then the process further comprises marking as the T point (T-peak) the stationary T point having the maximum value if the stationary T point with the minimum value is positive relative to the isoelectric line.
[0046] In an aspect 24 according to any one of the three preceding aspects, if the step of calculating said one or more stationary T points results in the identification of a plurality of stationary points, then the process further comprises marking as the T point (T-peak) the higher in absolute value between the stationary T point with the minimum value and the stationary T point with the maximum value if the stationary T point with the maximum value is positive relative to the isoelectric line and the stationary T point with the minimum value is negative relative to the isoelectric line.
[0047] In an aspect 25 according to any one of the preceding aspects, the electrocardiogram signal is a voltage signal. In an aspect 26 according to any one of the preceding aspects, the process comprises a procedure of aligning the electrocardiogram signal with an isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials, said isoelectric line coinciding with one or more segments of the electrocardiogram signal having zero voltage. In an aspect 27 according to the preceding aspect, the procedure of aligning the electrocardiogram signal comprises the steps of subdividing, by amplitude, one or more of said cardiac cycles into voltage sub-intervals, optionally comprised between 0.01 mV and 0.1 mV. In an aspect 28 according to the preceding aspect, the procedure of aligning the electrocardiogram signal further comprises the steps of counting the number of points making up the electrocardiogram signal of said one or more cardiac cycles, for each of said voltage sub-intervals. In an aspect 29 according to the preceding aspect, the procedure of aligning the electrocardiogram signal further comprises the steps of:
[0048] - determining the voltage sub-interval having the largest number of points,
[0049] - determining a voltage offset value of the voltage sub-interval having the largest number of points, optionally as a differential value between an arithmetic mean of the voltage values of the discrete points belonging to the interval with the largest number of points and a zero value,
[0050] - aligning the electrocardiogram signal with the isoelectric line as a function of the voltage offset value.
[0051] In an aspect 30 according to the preceding aspect, the procedure of aligning the electrocardiogram signal comprises at least one of the following steps of:
[0052] - comparing the number of points belonging to the voltage sub-interval with the largest number of points with the total number of points of one of said one or more cardiac cycles, optionally determining whether the ratio between the number of points belonging to the voltage sub-interval with the largest number of points and the total number of points of a respective one of said one or more cardiac cycles, is above 0.2,
[0053] - verifying that the mean voltage of the points belonging to the voltage sub-interval with the largest number of points is less than 0.5mV, optionally less than 0.1 mV, even more optionally zero.
[0054] In an aspect 31 according to any one of the preceding aspects, identifying the QRS complex (QRS) comprises the substeps of:
[0055] - filtering each of said one or more cardiac cycles of the electrocardiogram signal by means of a band-pass filter, optionally with a frequency comprised between 8Hz and 20Hz,
[0056] - calculating the absolute value of the filtered cardiac cycle.
[0057] In an aspect 32 according to any one of the preceding aspects, identifying the QRS complex (QRS) further comprises the sub-steps of determining, for each cardiac cycle of the electrocardiogram signal, a binary signal having:
[0058] o one or more main segments having a value greater than a QRS threshold value (QRS_Threshold), optionally comprised between 0.1 mV and 0.2mV,
[0059] o one or more secondary segments having a value less than the QRS threshold value (QRS_Threshold), optionally having a value of zero.
[0060] In an aspect 33 according to any one of the preceding aspects, the QRS complex (QRS) has:
[0061] - at least one negative segment representative of segments of the electrocardiogram signal having zero or negative values relative to the isoelectric line of the electrocardiogram signal having zero voltage,
[0062] - at least one positive segment representative of segments of the electrocardiogram signal having positive values relative to the isoelectric line.
[0063] In an aspect 34 according to either of the two preceding aspects, the binary signal is variable as a function of a current time (t). In an aspect 35 according to any one of the three preceding aspects, identifying the QRS complex (QRS) further comprises the steps of:
[0064] - measuring a time duration of each of said one or more main segments of the binary signal,
[0065] - comparing the measured time duration of said one or more main segments of the binary signal with a threshold duration value (QRSJnterval), optionally greater than 50ms,
[0066] - marking the QRS complex as the segment of the electrocardiogram signal corresponding to the main segment of the binary signal having a duration greater than the threshold duration value.
[0067] In an aspect 36 according to any one of the four preceding aspects, if a plurality of said one or more main segments of the binary signal have a duration greater than the threshold duration value, then the process envisages marking the QRS complex as the segment having the longest time duration of said one or more main segments of the binary signal with a duration greater than the threshold duration value. In an aspect 37 according to any one of the preceding aspects, the process comprises a step of determining a highest positive peak of the cardiac cycle of the electrocardiogram signal representative of an absolute maximum point of said cardiac cycle.
[0068] In an aspect 38 according to any one of the preceding aspects, the process comprises a procedure of calculating the negative curvature of the cardiac cycle of the electrocardiogram signal, said procedure of calculating the negative curvature, having the steps of:
[0069] - calculating, for every point of said series of discrete points below the isoelectric line, a curvature of the entire electrocardiogram signal,
[0070] - calculating a negative curvature threshold value as the arithmetic mean of said curvatures of the entire electrocardiogram signal,
[0071] In an aspect 39 according to any one of the preceding aspects, the process further comprises a procedure of calculating the positive curvature, having the steps of:
[0072] - calculating, for every point of said series of discrete points above the isoelectric line, a curvature of the entire electrocardiogram signal,
[0073] - calculating a positive curvature threshold value as the arithmetic mean of said curvatures of the entire electrocardiogram signal.
[0074] In an aspect 40 according to either of the two preceding aspects determining an S point of the QRS complex (QRS) comprises the steps of:
[0075] - determining one or more noteworthy S points among said series of discrete points of the cardiac cycle, where the first derivative of the cardiac cycle of the electrocardiogram signal is zero,
[0076] - comparing, for each of said one or more noteworthy S points, the curvature of the cardiac cycle in said one or more noteworthy S points with the negative curvature threshold value,
[0077] - identifying one or more minimum points between said one or more noteworthy S points in which the cardiac cycle has a curvature value greater than the negative curvature threshold value and / or having a negative voltage value.
[0078] In an aspect 41 according to the preceding aspect, determining an S point (S-peak) comprises the step of marking the S point (S-peak) as overlying the highest positive peak of the cardiac cycle of the electrocardiogram signal if the cardiac cycle, in none of said one or more noteworthy S points, has a curvature value greater than the negative curvature threshold value. In an aspect 41 according to the aspect 40, determining an S point (S-peak) comprises the step of marking as the S point (S-peak) the point between said one or more minimum points, adjacent to the highest positive peak. In an aspect 43 according to any one of the three preceding aspects, if identifying one or more minimum points between said one or more noteworthy S points determines the presence of a plurality of noteworthy S points downstream of the highest positive peak, determining an S point comprises the step of verifying that between the S point (S-peak) and the highest positive peak there is no presence of a further positive peak, representative of a relative maximum point of the cardiac cycle of the electrocardiogram signal with a curvature greater than the positive curvature threshold value.
[0079] In an aspect 44 according to the preceding aspect, if the step of verifying that between the S point (S-peak) and the highest positive peak gives a negative outcome, the process provides for marking as the S point (S-peak) a further of said one or more minimum S points and subsequently carrying out the steps of:
[0080] - reiterating the step of verifying that no further positive peak is present between the S point (S-peak) and the highest positive peak, and / or
[0081] - marking the S point (S-peak) as overlying the highest positive peak in the event that there are no further minimum points.
[0082] In an aspect 45 according to the preceding aspect, the step of verifying that no further positive peak is present between the S point (S-peak) and the highest positive peak envisages, as the first step, verifying the most negative peak between said one or more minimum points of said one or more noteworthy S points. In an aspect 46 according to any one of the preceding aspects, if identifying one or more minimum points between said one or more noteworthy S points determines the presence of a plurality of noteworthy S points upstream of the highest positive peak, determining an S point further comprises the steps of verifying the presence or absence of a further positive peak, representative of a relative maximum point of the cardiac cycle of the electrocardiogram signal with a curvature greater than the positive curvature threshold value, located upstream of the S point (S-peak). In an aspect 47 according to the preceding aspect, if the step of verifying the presence or absence of a further positive peak upstream of the S point (S-peak) gives a negative outcome, the process envisages marking as the S point (S-peak) a further of said one or more minimum S points. In an aspect 48 according to the preceding aspect, if the step of verifying the presence or absence of a further positive peak upstream of the S point (S-peak) gives a negative outcome, the process further envisages reiterating the step of verifying the presence or absence of a further positive peak upstream of the S point (S-peak). In an aspect 49 according to either of the two preceding aspects if the step of verifying the presence or absence of a further positive peak upstream of the S point (S-peak) gives a negative outcome, the process further envisages marking the S point (S-peak) as overlying the highest positive peak in the event that there are no further maximum points.
[0083] In an aspect 50 according to any one of the preceding aspects, if identifying one or more minimum points between said one or more noteworthy S points determines a first plurality of minimum points comprised between the highest positive peak and the terminal segment of a respective one of said one or more cardiac cycles and a second plurality of minimum points comprised between the highest positive peak and the initial segment of a respective one of said one or more cardiac cycles, determining an S point further comprises the steps of:
[0084] - selecting the most negative point among said first plurality of minimum points,
[0085] - selecting the most negative point among said second plurality of minimum points,
[0086] - determining one or more positive QRS peaks representative of relative maximum points of a respective one of said one or more cardiac cycles of the electrocardiogram signal,
[0087] In an aspect 51 according to the preceding aspect, determining the S point further comprises the steps of:
[0088] - calculating a curvature of the cardiac cycle of the electrocardiogram signal at said one or more positive QRS peaks,
[0089] - comparing the curvature of said one or more positive QRS peaks with a / the positive curvature threshold value.
[0090] In an aspect 52 according to the preceding aspect, determining the S point further comprises the steps of marking as the S point (S-peak) the most negative point of said one or more minimum points of the first plurality, if each of said one or more positive QRS peaks is less than the positive curvature threshold value. In an aspect 53 according to either of the two preceding aspects, determining the S point further comprises the steps of marking as the S point (S-peak) the minimum S point comprised between said positive QRS peak, optionally said relative maximum point, and the highest positive peak, if the step of determining one or more positive QRS peaks results in the identification of a single positive QRS peak, optionally a single relative maximum point.
[0091] In an aspect 54 according to any one of the preceding aspects, the process comprises a step of determining an R peak (R-peak) as a maximum point of the QRS complex (QRS). In an aspect 55 according to the preceding aspect, the R peak (R-peak) is comprised between an initial point of the QRS complex (QRS) of said series of discrete points and the S point (S-peak). In an aspect 56 according to either of the two preceding aspects, the step of determining an R peak (R-peak) comprises the sub-steps of:
[0092] - calculating a first derivative of the complex (QRS),
[0093] - determining, among said series of discrete points, one or more critical R points where the first derivative of the complex (QRS) is zero.
[0094] In an aspect 57 according to the preceding aspect, said one or more critical R points are above a / the isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials. In an aspect 58 according to either of the two preceding aspects, determining an R peak (R-peak) further comprises the sub-steps of:
[0095] - calculating a curvature of the complex (QRS) in said one or more critical R points,
[0096] - comparing the curvature of the complex (QRS) in said one or more critical R points with the positive curvature threshold value.
[0097] In an aspect 59 according to any one of the three preceding aspects, determining an R peak (R-peak) further comprises the sub-step of marking as the R peak (R-peak) the highest of said one or more critical R points where the curvature of the complex (QRS) is greater than the positive curvature threshold value. In an aspect 60 according to any one ofthe four preceding aspects, determining an R peak (R-peak) further comprises the sub-step of marking as the R peak (R-peak) a point between said one or more critical R points in which the QRS complex (QRS) has no curvature.
[0098] In an aspect 61 according to any one of the preceding aspects, the process comprises a step of determining, for each of said one or more cardiac cycles, a Q wave (Q-wave) representative of ventricular depolarisation. In an aspect 62 according to the preceding aspect, the Q wave (Q-wave) is comprised between the R peak (R-peak) and a / the initial point of the QRS complex (QRS). In an aspect 63 according to either of the two preceding aspects, the process comprises a step of determining a Q point (Q-peak) of said series of discrete points, representative of a maximum of the Q wave (Q-wave). In an aspect 64 according to the preceding aspect, determining the Q point (Q-peak) comprises the sub-steps of:
[0099] - determining, among said series of discrete points, one or more noteworthy Q points among said series of discrete points of the cardiac cycle, where the first derivative of the cardiac cycle of the electrocardiogram signal is zero, said noteworthy Q points being located upstream of the highest positive peak, optionally of the R peak (R-peak),
[0100] - comparing, for each of said one or more noteworthy Q points, the curvature of the cardiac cycle in said one or more noteworthy Q points with the negative curvature threshold value.
[0101] In an aspect 65 according to the preceding aspect, if at least one of said one or more noteworthy Q points is greater than the negative curvature threshold value, determining the Q point (Q-peak) comprises marking as the Q point (Q-peak) a point between said one or more noteworthy Q points adjacent to the R peak (R-peak).
[0102] In an aspect 66 according to either of the two preceding aspects, the process comprises determining an initial point of the Q wave (Q-onset) of said series of discrete points, a point between said one or more noteworthy Q points immediately preceding the Q point (Q-peak). In an aspect 67 according to any one of the three preceding aspects, if none of said one or more noteworthy Q points is greater than the negative curvature threshold value, determining the Q point (Q-peak) comprises marking as the Q point (Q-peak) a point between said one or more noteworthy Q points preceding the R peak (R-peak). In an aspect 68 according to any one of the four preceding aspects, the process comprises a step of marking as the Q point (Q-peak) a point of said series of discrete points coinciding with the highest positive peak, optionally with the R peak (R-peak). In an aspect 69 according to any one of the preceding aspects, determining an R' peak (R'-peak) comprises the sub-steps of:
[0103] - calculating a first derivative of a segment of the cardiac cycle comprised between the S point (S-peak) and an endpoint of the complex (QRS),
[0104] - determining, among said series of discrete points, one or more critical R' peaks where the first derivative of the complex (QRS) is zero.
[0105] In an aspect 70 according to the preceding aspect, said one or more critical R' peaks are above a / the isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials. In an aspect 71 according to either of the two preceding aspects, the step of determining an R' peak (R'-peak) comprises the sub-step of analysing the sign of the first derivative of a segment of the cardiac cycle comprised between the critical R' peak having the maximum value and the endpoint of the complex (QRS). In an aspect 72 according to the aspect 69 or 70, the step of determining an R' peak (R'-peak) further comprises the sub-step of analysing the sign of the first derivative of a segment of the cardiac cycle comprised between the critical R' peak having the maximum value and the point of said one ormore discrete points distant from said critical R' peak by a time interval comprised between 0.01s and 0.1s. In an aspect 73 according to any one of the four preceding aspects, determining an R' peak (R'-peak) further comprises the sub-step of marking as the R' peak (R'-peak) the critical R' peak having the maximum value if the sign of the first derivative of the segment of the cardiac cycle comprised between the R' peak with the maximum value and the endpoint of the complex (QRS) is negative. In an aspect 74 according to any one of the preceding aspects from 69 to 73, determining an R' peak (R'-peak) comprises the sub-step of marking the R' peak (R'-peak) as coinciding with the S point (S-peak).
[0106] In an aspect 75 according to any one of the preceding aspects, the process comprises calculating an angle of curvature comprised between a first straight line passing through the R' peak (R'-peak) and the J point (J-point) and a second straight line passing through the J point (J-point) and the T point (T-peak). In an aspect 76 according to the preceding aspect, the process comprises a step of determining whether the angle of curvature conforms to an acceptability criterion if the same angle of curvature is less than 200°, optionally comprised between 180° and 200°. In an aspect 77 according to the preceding aspect, the angle of curvature is calculated proceeding anticlockwise starting from the first straight line towards the second straight line. In an aspect 78 according to any one of the preceding aspects, the process comprises determining an intermediate point of the cardiac cycle of the electrocardiogram signal between the J point (J-point) and the T point (T-peak). In an aspect 79 according to the preceding aspect, the process comprises calculating an intermediate angle comprised between a primary straight line passing through the J point (J-point) and the intermediate point and a secondary straight line passing through the intermediate point and the T point (T-peak). In an aspect 80 according to the preceding aspect, the process further comprises a step of determining whether the intermediate angle conforms to an intermediate acceptability criterion, optionally if the intermediate angle is less than 180°. In an aspect 81 according to either of the two preceding aspects, the intermediate angle is calculated proceeding anticlockwise starting from the primary straight line towards the secondary straight line.
[0107] In an aspect 82 according to any one of the preceding aspects, the process comprises a step of determining, for each of said one or more cardiac cycles, a P wave (P-wave) as a segment of the electrocardiogram signal comprised between a / the final point of the T wave (T-endset) and an initial segment of the QRS complex (QRS) belonging to a temporally subsequent cardiac cycle. In an aspect 83 according to the preceding aspect, said P wave (P-wave) being representative of atrial depolarisation. In an aspect 84 according to either of the two preceding aspects, the process comprises determining a P point (P-peak) as a maximum point of the P wave (P-wave). In an aspect 85 according to the preceding aspect, determining the P point (P-peak) comprises the sub-steps of calculating, among said series of discrete points, one or more stationary P points wherein a first derivative of the segment of the electrocardiogram signal comprised between a / the final point of the T wave (T-endset), and an initial segment of the QRS complex (QRS) belonging to a temporally subsequent cardiac cycle is zero. In an aspect 86 according to the preceding aspect, if the step of calculating said one or more stationary P points results in the identification of a single stationary point, then the process envisages marking said single stationary point as the P point (P-peak). In an aspect 87 according to the aspect 85, if the step of calculating said one or more stationary P points results in the identification of a plurality of stationary points, then the process comprises determining a stationary P point with the maximum value and a stationary P point with the minimum value among the plurality of stationary P points.
[0108] In an aspect 88 according to any one of the three preceding aspects, if the step of calculating said one or more stationary P points results in the identification of a plurality of stationary points, then the process comprises marking as the P point (P-peak) the stationary P point with the minimum value if the stationary P point with the maximum value is negative in relation to a / the isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials. In an aspect 89 according to any one of the four preceding aspects, if the step of calculating said one or more stationary P points results in the identification of a plurality of stationary points, then the process comprises marking as the P point (P-peak) the stationary P point having the maximum value if the stationary P point with the minimum value is positive relative to the isoelectric line.
[0109] In an aspect 90 according to any one of the five preceding aspects, if the step of calculating said one or more stationary P points results in the identification of a plurality of stationary points, then the process comprises marking as the P point (P-peak) the higher in absolute value between the stationary P point with the minimum value and the stationary P point with the maximum value if the stationary P point with the maximum value is positive and the stationary P point with the minimum value is negative. In an aspect 91 according to any one of the preceding aspects the P wave (P-wave) is comprised between an initial point of the P wave (P-onset) of said series of discrete points and a final point of the P wave (P-endset) of said series of discrete points.
[0110] In an aspect 92 according to the preceding aspect, the step of determining the P wave (P-wave) comprises marking as the final point of the P wave (P-endset), a stationary P point immediately following the P point (P-peak). In an aspect 93 according to either of the two preceding aspects, the step of determining the P wave (P-wave) comprises marking as the initial point of the P wave (P-onset), a stationary P point immediately preceding the P point (P-peak). In an aspect 94 according to any one of the preceding aspects, the process comprises a step of determining a J wave (J-wave) representative of an irregularity in the cardiac cycle of the electrocardiogram signal at the J point (J-point). In an aspect 95 according to the preceding aspect, the process envisages determining a J wave (J-wave) if, in proximity to the J point (J-point), a double change of sign of a first derivative of the cardiac cycle of the electrocardiogram signal occurs. In an aspect 96 according to the preceding aspect, the J point (J-point) is at least partly overlying the J wave (J-wave).
[0111] In an aspect 97 according to any one of the preceding aspects, the process comprises a step of determining an epsilon wave (Epsilon-wave) representative of an irregularity in the cardiac cycle of the electrocardiogram signal after the J point (J-point). In an aspect 98 according to the preceding aspect, the epsilon wave (Epsilon-wave) is after the R' peak (R'-peak). In an aspect 99 according to either of the two preceding aspects, the process envisages determining an epsilon wave (Epsilon-wave) if, in a segment of the cardiac cycle following the J point (J-point), optionally following the R' peak (R'-peak), a double change of sign of a first derivative of the cardiac cycle of the electrocardiogram signal occurs. In an aspect 100 according to any one of the three preceding aspects, the J point (J-point) is distanced from the epsilon wave (Epsilon-wave).
[0112] In an aspect 101 according to any one of the preceding aspects, the process comprises a step of determining a delta wave (delta-wave) located in proximity to an initial segment of the QRS complex (QRS). In an aspect 102 according to the preceding aspect, the process envisages identifying the delta wave (delta-wave) at a change of concavity between the Q point (Q-peak) and the R point (R-peak) when Q is present, or between the start of the QRS and the R point. Inan aspect 103 according to the preceding aspect, the delta wave has a curvature greater than the positive curvature threshold value. In an aspect 104 according to any one of the preceding aspects, the process comprises the steps of:
[0113] - comparing a voltage value associated with the J point (J-point) with the isoelectric line of the electrocardiogram signal,
[0114] - diagnosing a heart condition a patient is affected by if the voltage value associated with the J point (J-point) is above or below the isoelectric line.
[0115] In an aspect 105 according to any one of the preceding aspects, the process that comprises a cardiac pathology of a patient comprising diagnosing a heart condition of a patient comprises diagnosing Brugada syndrome if the angle of curvature does not conform to an / the acceptability criterion. In an aspect 106 according to any one of the preceding aspects, the process that comprises diagnosing a heart disease of a patient further comprises diagnosing a STEMI acute coronary syndrome if the intermediate angle does not conform to the intermediate acceptability criterion.
[0116] In an aspect 107, there is envisaged a system for analysing an electrocardiogram trace comprising at least one control unit (50) configured to carry out the process of analysing at least one electrocardiogram trace relating to one or more leads of an electrocardiogram of a patient's heart in accordance with any one of the preceding aspects. In an aspect 108 according to the preceding aspect, the system comprises a plurality of electrodes configured to be positioned on a patient, each electrode of said plurality of electrodes being configured to generate an electrical signal representative of a cardiac activity of the patient. In an aspect 109 according to the preceding aspect, the control unit (50) is connected to the plurality of electrodes and configured to:
[0117] - determine one or more electrocardiogram signals as a function of each electrical signal generated by a respective electrode of said plurality,
[0118] - carry out the process according to any one of the preceding aspects from 1 to 106 on the basis of said one or more electrocardiogram signals.
[0119] In an aspect 110 according to the preceding aspect, the system comprises a memory (51) connected to the control unit (50) and configured to store the electrocardiogram signals determined by the control unit (50) itself. In an aspect 111 according to either of the two preceding aspects, the system comprises a screen (52) connected to the control unit (50) and configured to display the electrocardiogram signals determined by the control unit itself. In an aspect 112 according to any one of the five preceding aspects, the control unit (50) is configured to process said one or more electrocardiogram signals to obtain one or more lead signals, wherein each lead signal is a differential signal between two electrocardiogram signals. In an aspect 113 according to the preceding aspect, said one or more lead signals comprise bipolar lead signals of the limbs representative of a potential difference between two electrodes placed on the patient's limbs.
[0120] In an aspect 114 according to either of the two preceding aspects, said one or more lead signals comprise augmented unipolar lead signals representative of a potential difference between a central reference electrode and a further electrode on one of the patient's limbs. In an aspect 115 according to the preceding aspect, the central reference electrode is obtained as a mean of the voltage detected by electrodes located on the patient's limbs. In an aspect 116 according to any one of the four preceding aspects, said one or more lead signals comprise precordial lead signalsrepresentative of a potential difference between an electrode positioned on the patient's chest and the central reference electrode.
[0121] BRIEF DESCRIPTION OF THE DRAWINGS
[0122] Several embodiments and several aspects of the invention will be described hereinbelow with reference to the appended drawings, provided solely by way of illustration and thus not limiting, wherein:
[0123] - Figure 1 is a schematic view of an electrocardiographic system in accordance with the present invention,
[0124] - Figures 2 to 5 are block diagrams relating to steps of an electrocardiogram signal analysis process in accordance with the present invention,
[0125] - Figures 6 and 7 are respective graphs of an electrocardiogram signal with variable voltage,
[0126] - Figures 8-12, 14, 16, 18, 20, 22, 28A, 28B, 29, and 33 are block diagrams relating to steps of the process in accordance with the present invention, for the determination of respective points, peaks, angles or segments of an electrocardiogram signal,
[0127] - Figures 13, 15, 17, 19, 21, 23-27, 30-32, 34A and 34B show respective graphs of an electrocardiogram signal with voltage that is variable as a function of time, in which respective points and / or waves of a trace are shown,
[0128] - Figure 35 is a block diagram relating to steps of the process in accordance with the present invention, for the diagnosis of Brugada syndrome,
[0129] - Figure 36 is a block diagram relating to steps of the process in accordance with the present invention, for the diagnosis of a right bundle branch block,
[0130] - Figure 37 is a block diagram relating to steps of the process in accordance with the present invention, for the diagnosis of STEMI acute coronary syndrome.
[0131] DEFINITIONS AND CONVENTIONS
[0132] It should be noted that in the present detailed description, corresponding parts illustrated in the various figures are indicated with the same numerical references. The figures could illustrate the subject matter of the invention by means of non-scale representations; therefore, parts and components illustrated in the figure relating to the subject matter of the invention could exclusively regard schematic representations.
[0133] The electrocardiographic system described and claimed comprises / uses at least one control unit 50 designed to control operating conditions implemented by the electrocardiographic system itself and / or control the steps of the process claimed / described. The control unit 50 can be a single unit or be made up of a plurality of distinct control units depending on design choices and operating needs.
[0134] Control unit means a component of an electronic type which can comprise at least one of: a digital processor (CPU), a circuit of an analog type, or a combination of one or more digital processors with one or more circuits of an analog type. The control unit can be “configured’ or “programmed’ to carry out several steps: this can be achieved in practice with any means that allows the control unit to be configured or programmed. For example, in the case of a control unit comprising one or more CPUs and one or more memories, one or more programs can be stored in appropriate memory banks connected to the CPU or CPUs; the program or programs contain instructions which, when executed by the
[0135] CPU or CPUs, program or configure the control unit to carry out the operations described in relation to the control unit. Alternatively, if the control unit is / or comprises circuitry of an analog type, the circuit of the control unit can be designed to include circuitry configured, when in use, to process electrical signals in order to carry out the steps relating to the control unit. Parts of the process described here can be performed by means of a data processing unit, or control unit, technically replaceable with one or more computers conceived to execute a portion of a software or firmware program loaded onto a storage medium. The software program can be written in any programming language of a known type. The computers, if in a number of two or more, can be connected together by means of a data connection such that their computing powers may be shared in any manner; the computers themselves can thus be installed in positions that are also geographically distinct, thereby creating, by means of the aforesaid data connection, a distributed computing environment. The data processing unit, or control unit, can be a general-purpose processor configured to carry out one or more parts of the process identified in the present disclosure by means of the software or firmware program, or be an ASIC or dedicated processor or an FPGA, specifically programmed to carry out at least part of the operations of the process described herein. The storage medium can be non-transitory and can be internal or external to the processor, or control unit, or data processing unit, and can - specifically - be a memory located geographically remotely from the computer. The storage medium can likewise be physically divided into several portions, or in cloud form, and the software or firmware program can be stored in portions of memory geographically separated from one another.
[0136] DETAILED DESCRIPTION
[0137] The present invention relates to an electrocardiographic system 1 configured to sample electrocardiogram signals and carry out the steps of an electrocardiogram signal analysis process described below. As will be better described below, the process is configured to perform an analysis of electrocardiogram (ECG) traces of a patient in order to identify morphological components, i.e. waves (for example P, QRS, T waves), intervals and segments, which describe the various phases of the heart's electrical activity.
[0138] The electrocardiographic system 1 comprises a plurality of electrodes 2, intended to be positioned on a patient's body. Figure 1 shows three electrodes for the sake of simplicity of representation; however, the electrocardiographic system can comprise twelve or more electrodes. Each electrode 2 is configured to read an electrical signal that represents a cardiac activity of the patient. The electrocardiographic system 1 can comprise a control unit 50, connected to the electrodes 2 in order to receive the signals generated by the latter and generate one or more electrocardiogram traces. The electrocardiogram signal processed by the control unit 50 can be, for example, a digital signal that is variable as a function of time and represented by a series of discrete points. The control unit 50 can be further configured to carry out the process of analysis described below and identify cardiac waves, intervals and segments that characterise the various phases of the heart's electrical activity.
[0139] Furthermore, the electrocardiographic system 1 can comprise at least one memory 3, connected to or incorporated into the control unit 50, which enables the acquired electrocardiogram signals to be stored for subsequent analyses. The electrocardiographic system 1 can further comprise a screen 4, likewise connectable to the control unit 50 in order to display the electrocardiogram traces in real time. The control unit 50 can be further configured to display, on thescreen 4, indications or notifications for the operator relating to the identification of characteristics of the electrocardiogram signal. In a variant embodiment, the electrocardiographic system can also be configured to diagnose one or more heart conditions based on the identification of characteristics of the electrocardiogram trace conforming to an acceptability criterion and / or an intermediate acceptability criterion detailed further below. Information relating to the diagnosis performed can be displayed on the screen 4 or sent, via the system 1, to a remote device of a healthcare professional.
[0140] The control unit 50 can further be configured to process the electrocardiogram signals in order to obtain lead signals. Every lead signal is the result of a potential difference measured between two or more electrodes. The leads can include:
[0141] - bipolar limb leads (DI, Dll, Dill), which measure the potential difference between two electrodes 2 placed on the patient's limbs. For example, DI measures the difference between the right arm and left arm, Dll between the right arm and left leg, and Dill between the left arm and left leg,
[0142] - augmented unipolar limb leads (aVR, aVL, aVF), which measure the potential difference between one electrode on a limb and a central reference electrode, obtained as the mean of the voltage measured by the electrodes on the other two limbs. The leads include aVR (right arm), aVL (left arm), and aVF (left leg),
[0143] - precordial leads (V1-V6), which measure the potential difference between an electrode positioned on the patient's chest and the central reference electrode. These leads provide detailed information about the electrical activity in the different areas of the heart.
[0144] The control unit 50 can be further configured to analyse the electrocardiogram signals acquired from each of the abovedescribed leads, in order to identify the various components of the electrocardiogram signal and obtain a complete view of the heart's electrical activity, thereby enabling an accurate assessment of heart rate, conduction and the presence of any anomalies.
[0145] It should be noted that the electrocardiogram signal analysis process illustrated below can be directly implemented by the described electrocardiographic system 1, or, in an alternative configuration, it can be carried out by software installed on a processor of a healthcare professional other than the electrocardiographic system, configured to receive a previously sampled electrocardiogram signal.
[0146] The electrocardiogram signal analysis process of the present invention and implemented by the system 1, in a first step, can envisage a pre-processing of the electromyographic signal focused on reducing noise and artefacts which could compromise signal accuracy (figure 2). The raw signal, for example sampled by the electrocardiographic system 1 , can be subject to various types of interference caused by body movements or electrode contact problems, or caused by electrical interference at medium-high frequencies (for example higher than 50Hz). Other artefacts include muscle noise (EMG), which derives from surrounding muscle activity, and movements of the electrodes, which can cause distortions. In order to attenuate such disturbances, the process can envisage a step of filtering the sampled electrocardiogram signal in the frequency band comprised between 0.5Hz-50Hz, thus eliminating frequencies outside the band (step 100 in figure 2).
[0147] Once the signal filtering step has been carried out, the process can envisage a step 101 of detecting an isoelectric line of the electrocardiogram signal under examination, representative of a zero electric potential of the signal and usableas a reference in order to correct any misalignments of the electrocardiogram signal. In detail, the isoelectric line represents the moment of the heartbeat in which the electric potential is absent, i.e. when no cardiac action potentials are detected. However, due to interference or residual noise in the signal, segments of the electrocardiogram signal that should show zero voltage have a voltage value other than zero.
[0148] In this context, the process can envisage an electrocardiogram signal alignment procedure, indicated in the block diagram in figure 2 by the reference number 102. For the sake of simplicity of illustration, the steps of the electrocardiogram signal alignment procedure will be detailed below with reference to the block diagram in figure 3 and the graph shown in figure 4. Following the description of said alignment procedure, the illustration of the remaining steps related to the pre-processing of the electrocardiogram signal will be resumed, again making reference to figure 2.
[0149] In one example, the electrocardiogram signal alignment procedure can comprise a first step of subdividing the electrocardiogram signal into intervals of a predetermined amplitude (step 110 in figure 3). Making reference to the graph in figure 4, which shows an electrocardiogram signal with voltage that is variable as a function of time, this subdivision of the electrocardiogram signal envisages dividing the signal, by amplitude, into voltage sub-intervals, optionally comprised between 0.01 mV and 0.1 mV, each representing a range of voltages. The alignment procedure can further comprise a step 111 of counting the number of points of the ECG signal which fall within each voltage subinterval. The voltage interval with the maximum concentration of points is subsequently determined as the zero-voltage segment of the electrocardiogram signal (step 112), as this represents the most frequent voltage value.
[0150] The alignment procedure can also comprise performing two check steps, summarily indicated in figure 3 by the reference number 113, to verify that the voltage sub-interval identified actually corresponds to the isoelectric line. A first step comprises comparing the number of points in the voltage sub-interval with the maximum concentration with the total number of points of the sampled signal and, subordinately, determining whether the ratio between the number of points belonging to the voltage sub-interval with the maximum concentration and the total number of points of the electrocardiogram signal is greater than 0.2. A second step envisages verifying that the mean voltage of the points belonging to the voltage sub-interval with the largest number of points is zero.
[0151] Subsequently, the process can envisage determining a voltage offset value (step 114), calculated as the difference between the arithmetic mean of the voltage values of the discrete points within the sub-interval with the maximum concentration and a zero value. Finally, the signal (in particular the ECG) is realigned with the isoelectric line on the basis of the determined voltage offset value, thus correcting any misalignment and bringing the signal back to the correct zero potential reference (step 115). In this context, see for example the graph in figure 4, where the filtered electrocardiogram signal is represented by a broken line (identified in figure 4 as ECG Clean). In this case, the signal is shifted vertically to align the points that should have a zero potential with the isoelectric line 300. The result of the realignment is shown in figure 4 as a solid line identified as ECG Isoelectric.
[0152] Making reference again to the block diagram in figure 2, the pre-processing of the signal can also comprise a step 103 of identifying a QRS complex indicative of a segment of the electrocardiogram signal representative of ventricular depolarisation. As shown in the block diagram in figure 5, a first sub-step relating to the identification of the QRS complex can envisage isolating the frequency components pertaining to the QRS complex by filtering the
[0153] electrocardiogram signal by means of a band-pass filter, with cut-off frequencies comprised between 8Hz and 20Hz (step 120).
[0154] Once the signal has been filtered, the process can envisage dividing the electrocardiogram signal into one or more cardiac cycles, each of which representing a single cardiac cycle (step 121). Subsequently, the process can envisage generating a binary signal (step 122) having main segments marked with a unitary value and secondary segments identified with a zero value. In one example, the main segments represent segments of the cardiac cycle of the signal that exceed a QRS threshold value (identified in the figure as QRS_Threshold) comprised between 0.1 mV and 0.2mV, whilst the secondary segments are representative of segments of the cardiac cycle of the signal below the QRS threshold value. A step 125 of comparing segments or individual points of the cardiac cycle with the QRS threshold value is illustrated in figure 5. This comparison makes it possible to determine both the main segments (step 123) and the secondary segments (step 124) of the electrocardiogram signals, i.e. the characteristic portions of the lead signal, distinguished on the basis of clinical and morphological relevance.
[0155] The process proceeds by identifying the QRS complex internally between the main segments (i.e. the fundamental components of the electrocardiogram signal, present under the normal and clinically most significant conditions) of the generated binary signal. In this context, the process envisages measuring a time duration of each main segment and comparing it with a threshold duration value (indicated in figure 5 by the reference term QRSJnterval), optionally greater than 50ms (step 126). If the main segment is greater than the threshold duration value, it is marked as the QRS complex (step 127); otherwise, there is no identification of the QRS complex (step 128). It should be noted that the comparison step 127 envisages of iteratively comparing all the segments of the cardiac cycle with the threshold duration value. If a number of main segments of the binary signal with a duration greater than the threshold are present, the segment with the longest duration is identified as the QRS complex.
[0156] The step of pre-processing the signal can also envisage a step of detecting the highest positive peak in each cardiac cycle of the electrocardiogram signal and representative of an absolute maximum point of a respective cardiac cycle (see step 104 in figure 2). Once the QRS complex has been identified and the highest positive peak determined, the process goes on to identify the components of the electrocardiogram signal and analyse the morphological characteristics thereof in accordance with what is detailed below.
[0157] Before the steps of the electrocardiogram signal analysis process are illustrated, the morphology of an electrocardiogram signal will be described in general terms and the main components thereof briefly introduced. Making reference to figure 6, the electrocardiogram signal (ECG) is composed of various components that appear in a specific order, corresponding to the phases of the cardiac cycle. The first of these is the P wave, which represents atrial depolarisation and marks the start of contraction of the atria. One subsequently observes the QRS complex, which reflects the depolarisation of the ventricles. Within this complex, the Q wave indicates the depolarisation of the interventricular septum, whilst the R wave corresponds to the main ventricular depolarisation, and the S wave represents the depolarisation of the basal portions of the ventricles. In some situations, a further positive wave called R' can also appear, which occurs after the S wave, signalling a second phase of depolarisation in certain physiological or pathological conditions. The QRS complex is followed by the T wave, which is an expression of ventricular repolarisation, i.e. the return of the ventricles to a state of rest. In some cases, a U wave can also appear, reflecting alate repolarisation of the ventricles. Located between the QRS complex and the T wave is the J point (shown in figure 27), which marks the end of ventricular depolarisation. Together these components, arranged according to their order in time, enable a reading of the cardiac cycle.
[0158] Hereinafter in the description, reference will further be made to the calculation of the curvatures of the electrocardiogram signal, to be compared with one or more threshold values. It should be noted that the term "curvature of the electrocardiogram signal in a point” means the measurement indicating how much a curve (for example the ECG signal) deviates from a straight line in that point. Mathematically, it is defined as the rate of change of the direction of the tangent to the curve relative to the length of the arc. In other words, the curvature describes how fast the curve changes direction in a specific point: the greater the curvature, the more accentuated the deviation from the rectilinear direction.
[0159] Depending on whether the characteristic to be identified is positive or negative relative to the isoelectric line, the process can envisage carrying out a procedure of calculating the negative curvature of a cardiac cycle of the electrocardiogram signal and / or a procedure of calculating the positive curvature. The procedure of calculating the negative curvature has the steps of:
[0160] - calculating, for every point of the electrocardiogram signal below the isoelectric line, a curvature of the entire electrocardiogram signal,
[0161] - calculating a negative curvature threshold value as the arithmetic mean of said curvatures of the entire electrocardiogram signal,
[0162] Similarly, the procedure of calculating the positive curvature has the steps of:
[0163] - calculating, for every point of said series of discrete points above the isoelectric line, a curvature of the entire electrocardiogram signal,
[0164] - calculating a positive curvature threshold value as the arithmetic mean of said curvatures of the entire electrocardiogram signal,
[0165] Hereinafter, the steps of the algorithm dedicated to identifying the components of the electrocardiogram signal will be presented following the sequential order of the process itself. However, this sequence is not obligatory and performing the steps of the algorithm according to an alternative order can be envisaged.
[0166] With reference to figure 7, the process can envisage a step of determining an S point, identified with S-peak, belonging to an S wave defined within the QRS complex. Hereinafter in the description, reference will further be made to the block diagram in figure 8 in order to illustrate the various steps leading to the identification of the S point. The process can initially envisage taking as reference the highest positive peak of the cardiac cycle of the electrocardiogram signal, as the central search reference, which, in the graph in figure 7 has been identified as the R point. Subsequently, the process envisages determining one or more noteworthy S points (step 130 in figure 8), defined as the points in which the first derivative of the electrocardiogram signal is zero. Such points are then divided into two groups: the ones located downstream and the ones upstream of the highest positive peak.
[0167] Considering that, in a normal cardiac cycle, the S point has a negative value relative to the isoelectric line, in each noteworthy S point, the process envisages calculating a curvature of the electrocardiogram signal (step 131) and comparing it with the negative curvature threshold value (step 132). Only the noteworthy S points that show a curvaturegreater than the negative curvature threshold value and a negative voltage value will be considered as possible candidates for identifying the S point, whereas the noteworthy S points with a curvature less than the threshold value will be discarded (step 133). In some morphologies, it might not be sufficient that the electrocardiogram signal meets such requirements in order to unambiguously identify the S point, since also a Q wave could have a curvature greater than the negative curvature threshold value. In this context, to avoid an incorrect identification of the S point, the process can envisage carrying out four independent procedures based on cases subsequently illustrated and identified in figure 8 with the references 134 to 137.
[0168] In a first case, illustrated in the block diagram in figure 9, if none of the noteworthy S points has a curvature greater than the negative curvature threshold value and has a negative voltage value, the electrocardiogram signal will lack an S curve. Considering the consequentiality of the process, this can entail fictitiously marking the S point as coinciding with the highest positive peak or as the noteworthy S point adjacent to the highest positive peak (step 138).
[0169] In a second case shown in figure 10, if the presence of noteworthy S points that satisfy the abovementioned conditions is detected and they are located only after the highest positive peak, the process can envisage initially a step 139 of marking the highest positive peak as the R peak (identified in the figures as the R-peak). Subsequently, the process can comprise performing an iterative step to verify which of the noteworthy S points following the highest positive peak that satisfy the above-described conditions can be marked as the S point. A first step can envisage marking as the S point the lowest of the noteworthy S points identified, i.e. the one having the lowest voltage value (step 140), and verifying whether between that noteworthy S point and the highest positive peak there is no presence of a further positive peak, which would represent a relative maximum of the cardiac cycle of the electrocardiogram signal with a curvature greater than the positive threshold value (step 141). If this verification gives a positive outcome, the iterative step will terminate and the S point will remain marked as the lowest of the noteworthy S points. If, by contrast, the presence of a further positive peak is determined between the noteworthy S point identified and the highest positive peak, the process will be repeated, taking into consideration the most negative noteworthy S point among the remaining noteworthy S points (step 142) and reiterating the verification step to ensure the absence of further positive peaks. However, if there are no more noteworthy S points remaining, or if none of them satisfies the required conditions (noteworthy S point with a curvature greater than the negative curvature threshold value and a negative voltage value), the S point will be marked as coinciding with the highest positive peak, i.e. as coinciding with the R peak.
[0170] In a third case and with reference to figure 11, if the presence of noteworthy S points which have a curvature greater than the negative curvature threshold value and a negative voltage value is detected only before the highest positive peak, the process can envisage initially a step 145 of marking the highest positive peak as the R' peak (identified in the figures as the R'-peak). Subsequently, the process comprises performing an iterative step to verify which of the noteworthy S points preceding the highest positive peak that satisfy the above-described conditions can be marked as the S point. A first step can envisage marking as the S point the lowest of the noteworthy S points identified, i.e. the noteworthy S point having the lowest voltage (step 146) and verifying whether a further positive peak with a curvature greater than the positive threshold value is present upstream of the S point (step 147). If this verification gives a positive outcome, the iterative step will terminate and the S point will remain marked as the lowest of the noteworthy S points; otherwise, the process will be repeated, taking into consideration the most negative noteworthy S point among theremaining noteworthy S points (step 148) and reiterating the verification step 147 to ensure the absence of further positive peaks. However, if there are no more noteworthy S points remaining, or if none of them satisfies the required conditions (noteworthy S point with a curvature greater than the negative curvature threshold value and a negative voltage value), the S point will be marked as coinciding with the highest positive peak, i.e. as coinciding with the R' peak.
[0171] In afourth case shown in figure 12, the process focuses on managing the eventuality that may lead to the determination of noteworthy S points where the electrocardiogram signal has a curvature greater than the negative curvature threshold value and a negative voltage value, and which are located both upstream and downstream of the highest positive peak.
[0172] The process envisages initially determining two distinct groups of minimum points. A first group is made up of minimum points located between the highest positive peak and the terminal segment of the cardiac cycle under examination, whereas a second group comprises the minimum points comprised between the highest positive peak and the initial segment of the same cardiac cycle (steps 150 and 151 of the block diagram in figure 12). Once this distinction has been established, the process proceeds by selecting the most negative point within each of these two groups, i.e. the point having a minimum voltage value (step 152). At this point, one or more positive QRS peaks are identified (step 153), which represent the relative maximum points of the cardiac cycle of the electrocardiogram signal. In each of these positive QRS peaks, a calculation is made of the curvature of the cardiac cycle of the signal, to be compared with the positive curvature threshold value.
[0173] If all the positive QRS peaks have a curvature that is less than the negative curvature threshold value, the S point (S-peak) will be marked as the most negative point of the first group of minimum points, i.e. the one located downstream of the highest positive peak. If, by contrast, the process detects a single positive QRS peak, then the S point (S-peak) will be identified as the minimum point located between that positive QRS peak and the highest positive peak. The step of marking the S point is indicated overall in figure 12 by the reference number 154.
[0174] The process can further envisage an R wave analysis for the identification of the R peak, indicated in figure 13 by the reference R-peak, as a maximum point of a positive portion of the QRS complex. With reference to the block diagram in figure 14, the process may envisage calculating the first derivative of the QRS complex to determine critical R points (step 160), or all the points of the QRS complex where the first derivative of the same QRS complex is zero. Among these critical points, searched for within a cardiac cycle temporally preceding the identified S point, only those above the isoelectric line, i.e. having a positive electric potential, will be selected. Thereafter, the process envisages calculating a curvature of the QRS complex in each of the critical R points (step 161) and subsequently comparing the curvature values obtained with a positive curvature threshold value. This step makes it possible to distinguish the relevant critical points, i.e. the ones in which the curvature of the signal is sufficiently pronounced to indicate a significant peak. Once the curvature analysis has been completed, the process can have a step 162 of determining the R peak on the basis of two possible conditions. If one or more critical R points have a curvature greater than the mean curvature of the QRS complex, the R peak (R-peak) will be identified as the critical point having the maximum value among them. If, by contrast, there are no critical points that satisfy this condition, the R peak (R-peak) will be marked as a critical R point in which the curvature of the QRS complex is zero, a condition indicating the lack of an Rpeak. In the latter case, the cardiac cycle of the electrocardiogram signal under examination will have a QS complex and not a QRS complex.
[0175] The process can further comprise a step of determining a Q wave of the QRS complex (identified in figure 15 as Q-wave). The Q wave (Q-wave) represents the first negative curve within the QRS complex and is representative of a phase of the cardiac cycle associated with ventricular depolarisation. The Q wave is identified between the R peak and an initial point of the QRS complex. In particular, the process involves a series of steps aimed at identifying the Q peak (Q-peak), which represents a maximum point of the Q wave. In this context, the process can envisage initially identifying one or more noteworthy Q points, located upstream of the highest positive peak, where the first derivative of the cardiac cycle of the electrocardiogram signal is zero (step 170 of the block diagram in figure 16). The process can subsequently envisage calculating a curvature of the cardiac cycle of the electrocardiogram signal at the noteworthy Q points (step 171) and, subordinately, comparing the calculated curvature of the cardiac cycle with the negative curvature threshold value. Subsequent steps of the process are dedicated to marking the Q point according to the following cases identified overall in the block diagram in figure 16 by the reference number 172. In a first case, if at least one noteworthy Q point greater than the negative curvature threshold value is detected, the process envisages marking as the Q point the noteworthy Q point adjacent to, optionally immediately preceding, the R peak. In this scenario, the process can also envisage determining an initial point of the Q wave (identified as Q-onset in figure 15), as the noteworthy Q point immediately preceding the Q point just marked.
[0176] In asecond case, if none of said noteworthy Q points isgreaterthan the negative curvature threshold value, the process can envisage marking as the Q point the noteworthy Q point immediately preceding the R peak.
[0177] A third case takes into consideration the eventuality of an absence of the R wave, such as to result in the detection of a QS complex. In such a scenario, the process can envisage fictitiously marking the Q point as coinciding with the highest positive peak, optionally with the R peak.
[0178] As previously mentioned, the process can also envisage identifying the R' peak, located immediately following the S point, within the QRS complex and representative of a second ventricular depolarisation. The R' peak, identified in figure 17 as the refence R'-peak, corresponds to the last positive peak of the QRS complex and is possibly only present in the right leads (aVR, V1, V2, V3). With reference to the block diagram in figure 18, the process can envisage calculating the first derivative of the cardiac cycle comprised between the S point and the endpoint of the QRS complex, resulting in the identification of one or more critical R' peaks in which the derivative is zero (step 180). Among the critical R' peaks, only the ones above the isoelectric line are taken into consideration for the identification of the R' peak. Subsequently, the process can envisage a step of analysing the sign of the first derivative in two possible intervals (step 181). A first interval is comprised between the critical R' peak having the maximum value and the endpoint of the QRS complex, whereas a second interval can be defined between the critical R' peak with the maximum value and a further point of the electrocardiogram signal, situated at a temporal distance ranging between 0.01 and 0.1 seconds from the critical R' peak with the maximum value. If the sign of the first derivative is negative in the segment comprised between the critical R' peak with the maximum value and the endpoint of the QRS complex, the R' peak will be identified as coinciding with the critical R' peak having the maximum value. If, by contrast, that condition is not satisfied, the R'peak will be fictitiously marked as coinciding with the previously determined S point. It should be noted, finally, that in figure 18, the steps of determining the R' peak have been indicated overall by the reference number 182.
[0179] The process can further envisage determining a T point belonging to a T wave, identified respectively in figure 19 as T-wave and T-peak. The T wave represents the phase of the cardiac cycle corresponding to ventricular repolarisation and is characterised by a negative concavity in all the leads, with the exception of aVR, where it can have a different configuration. Normally, the T wave has an asymmetric profile, with an ascending phase that is slower than the descending one and a T-peak amplitude of less than 0.5mV in the peripheral leads and 1.5mV in the precordial leads. The search for the T wave takes place within the secondary segments, i.e. the segments with a value of zero, of the binary signal previously calculated for the identification of the QRS complex. The search for the T wave thus takes place outside the high frequency zone of the QRS complex. In this scenario, the process can envisage initially determining the T point and subsequently identifying an initial point of the T wave (indicated in figure 19 as T-onset) and a final point of the T wave (indicated in figure 19 as T-endset).
[0180] With reference to the block diagram in figure 20, the process can envisage a step 190 of calculating one or more stationary T points, i.e. points in which the first derivative of the electrocardiogram signal is zero. The process further envisages excluding the points located on the isoelectric line, since the T point cannot have zero voltage. If a single stationary point is identified, it is directly marked as the T point. If, by contrast, a plurality of stationary T points is determined, the process envisages determining a stationary T point with the maximum value and a stationary T point with the minimum value among the plurality of stationary T points calculated (step 191). Depending on the voltage values assumed by the stationary T point with the maximum value and by the stationary T point with the minimum value, the T point is determined (step 192) in accordance with the following cases. If the stationary T point with the maximum value is negative relative to the isoelectric line, the T point will be marked as coinciding with the stationary T point with the minimum value. In a second case, if the stationary T point with the minimum value is positive relative to the isoelectric line, the process will envisage marking the T point as coinciding with the stationary T point having the maximum value. Finally, in a third case, if the stationary T point with the maximum value is positive relative to the isoelectric line and the stationary T point with the minimum value is negative relative to the isoelectric line, the process will envisage marking the T point as coinciding with the higher in absolute value between the stationary T point with the minimum value and the stationary T point with the maximum value.
[0181] Subsequently, the process can also comprise identifying the initial point of the T wave (T-onset) and the final point of the T wave (T-endset). As regards the final point of the T wave, the process can envisage a step 193 of marking the stationary T point immediately following the T point as the final point of the T wave (T-endset).
[0182] As regards the initial point of the T wave, the process can envisage the identification thereof, in a step 194, only downstream of the identification of a J point representative of the passage between a phase of ventricular depolarisation (represented by the QRS complex) and a start of ventricular repolarisation represented by the T wave. In particular, the process can envisage analysing a JT segment, comprised between the J point (identified in figure 27 as the J-point) and the T point, for the purpose of identifying a change of concavity. More specifically, the process can envisage calculating a second derivative of the JT segment and, subordinately, analysing the sign of the second
[0183] derivative of the JT segment at the T point, in order to determine a concavity of the electrocardiogram signal at the T point.
[0184] A second step of the process envisages calculating one or more critical T points where the second derivative of the electrocardiogram signal is zero, not defined or inexistent. Subsequently, the process can envisage analysing the sign of the second derivative of the JT segment at each critical point calculated to determine a concavity of the electrocardiogram signal to be compared with the concavity of the electrocardiogram signal at the previously calculated T point. The process envisages marking as the initial point of the T wave the critical T point where the curvature of the JT segment has maximum absolute value and concavity opposed to that detected at the T point.
[0185] With reference to figure 21, the process can also comprise a step of identifying a P wave (as indicated) representative of a phase of atrial depolarisation in the cardiac cycle, which takes place between the final point of the T wave (T-endset) and the start of a QRS complex (QRS) belonging to a temporally subsequent cardiac cycle.
[0186] In one example, the process can be configured to detect the P wave within a time interval delimited by the final point of the T wave (T-endset) and a further limit point at a distance of 80ms both from the same final point of the T wave and from the start of the QRS complex of the next cardiac cycle. This constraint makes it possible to avoid a false interpretation of any U wave as the P wave, consequently improving the entire identification process. The steps of the process involving the identification of the P-curve are similar to those carried out in relation to the identification of the T-curve, providing initially for the identification of the P peak (P-peak), which is followed by the determination of an initial point of the P wave (P-onset) and a final point of the P wave (P-endset). With reference to the block diagram in figure 22, the process can envisage a step 200 of calculating one or more stationary P points wherein the first derivative of the segment of the electrocardiogram signal is zero. As mentioned for the identification of the T-curve, the points on the isoelectric line are excluded as they are representative of non-detectable electric potentials and are therefore incompatible with the P peak, which necessarily has a positive voltage. If a single stationary P point is identified, it will be marked as the P peak. If, by contrast, a plurality of stationary P points is determined, the process can envisage a step 201 of determining a stationary P point with the maximum value and a stationary P point with the minimum value among the plurality of stationary P points calculated. Depending on the voltage values assumed by the stationary P point with the maximum value and by the stationary P point with the minimum value, the P peak is determined, in a step 202, in accordance with the following cases. If the stationary P point with the maximum value is negative relative to the isoelectric line, the P peak will be marked as coinciding with the stationary P point with the minimum value. In a second case, if the stationary P point with the minimum value is positive relative to the isoelectric line, the process will envisage marking the P peak as coinciding with the stationary P point having the maximum value. Finally, in a third case, if the stationary P point with the maximum value is positive relative to the isoelectric line and the stationary P point with the minimum value is negative relative to the isoelectric line, the process will envisage marking the P peak as coinciding with the higher in absolute value between the stationary P point with the minimum value and the stationary P point with the maximum value.
[0187] Subsequently, the process can also comprise a step 203 of identifying the initial point of the P wave (P-onset) and a step 204 of identifying the final point of the P wave (P-endset), respectively as the stationary points immediately preceding and following the P peak. If the P wave is not present and, therefore, there are no stationary P points havinga non-zero voltage, the P peak, the initial point of the P wave and the final point of the P wave will be marked as the first stationary point preceding the initial point of the Q wave (Q-onset) that is situated on the isoelectric line.
[0188] With reference to figure 23, the process can also envisage identifying a U wave representative of a late repolarisation of ventricular muscle fibres. In one example, the process can envisage identifying the U wave within a time interval comprised between the endpoint of the T wave (T-endset) and a terminal point at a distance of 80 milliseconds from the endpoint of the T wave. Within this time interval, the process can exclusively envisage an analysis of the secondary segments of the binary signal previously calculated for the determination of the QRS complex, thus excluding the areas characterised by high frequency, typical of the QRS itself. If, within this interval, no points with a derivative of zero are detected, the U wave is considered absent. Conversely, the U wave or, more specifically, a maximum point thereof, can be marked in one of those points with a derivative of zero.
[0189] The process can further comprise the identification of at least one of: a J wave, shown in figure 24, an epsilon wave shown in figures 25A and 25B and a delta wave shown in figure 26. Such waves can be a symptom of irregularities that manifest themselves in specific segments of the electrocardiogram signal, each with distinctive characteristics tied to particular phases of the cardiac cycle.
[0190] The J wave (J-wave) is determined in proximity to the J point (J-point), representing an irregularity that appears during the transition between the end of the QRS complex and the start of an ST segment comprised between the S point and the T point. The presence of the J wave is identified by the occurrence of a double change of sign of the first derivative of the electrocardiogram signal near the J point, indicating a reversal in the direction of the trace which suggests a discontinuity or electrical variation characteristic of some specific heart diseases or conditions. The J wave can thus at least partly overlap the J point, sharing a common portion of the trace.
[0191] As regards the epsilon wave (Epsilon-wave), it can be located after the J point, and in some cases after the R' peak (R'-peak), if present. The identification of the epsilon wave is likewise based on the observation of a double change of sign of the first derivative of the electrocardiogram signal in the cardiac cycle following the J point. This characteristic suggests the presence of a deflection or undulation which can be indicative of anomalies in ventricular conduction or structural alterations of the myocardium. The distance between the J point and the epsilon wave may vary, allowing the epsilon wave to be identified also in segments of the trace that are at a certain distance from the J point.
[0192] Finally, the delta wave (Delta-wave) can be located in proximity to the initial segment of the QRS complex. The delta wave reflects an anomalous curvature of the signal at the start of ventricular depolarisation, generally identified between the Q point (Q-peak) and the R point (R-peak), when the Q point is present. In the absence of a Q point, the delta wave is identified between the start of the QRS and the R point. The distinctive characteristic of the delta wave is a curvature greater than the positive curvature threshold value, which signals a pronounced change in the morphology of the electrocardiogram signal that can be correlated to ventricular pre-excitation syndromes.
[0193] It should be noted that the identification process detailed above is related to the identification of waves and complexes. However, the process can also provide for the identification of specific points characterising the sampled electrocardiogram trace (for example the J point) and which provide additional information on the transitions between the various phases of the cardiac cycle.
[0194] In one example, the process for identifying the J point (J-point) envisages searching for the J point itself in a segment of the electrocardiogram signal comprised between the QRS complex and the T wave. In one example and with reference to the block diagram in figure 28A, the process can comprise a step 210 of determining a J point search segment representative of a segment of the electrocardiogram signal comprised between the S point and the T point if the absence of the R' peak is verified. Otherwise, if the presence of the R' peak is detected, the process envisages narrowing the J point search segment to between the R' peak and the T point. This step can make it possible to avoid an incorrect identification of the J point itself. Within the ST segment (or R'T segment), the process can envisage calculating one or more noteworthy J points, where a change in the concavity of the electrocardiogram signal is recorded (step 211). The calculation of the noteworthy J points envisages calculating a second derivative of the J point search segment, identifying the points wherein said derivative is equal to zero, does not exist or is not defined, and verifying the presence of a change of sign of the derivative itself, which indicates the actual change of concavity.
[0195] In the event that a single noteworthy J point representative of an inflection point is identified, the process can envisage marking it as the J point. If, by contrast, a number of noteworthy J points are identified, all representative of respective inflection points, the process can envisage further criteria for determining which of them represents the actual J point. In particular, one can choose the point in which the first derivative of the J point search segment is equal to zero or the point with the maximum curvature, situated between the S point (or R' peak, if present) and the T point. Both options dedicated to identifying the J point have been indicated in figure 28A by the reference number 212.
[0196] The process can further comprise one or more steps dedicated to analysing the quality of the characteristics identified in the electrocardiogram signal. It should be noted, in fact, that the sensitivity and specificity of the process can depend closely on the accuracy with which the characteristic parameters of the trace were identified. For this reason, at the end of the initial processing for the identification of the components, the process can envisage performing quality control steps divided into two distinct steps, which aim to assess and improve the validity of the results obtained.
[0197] The first quality control step is focused on the quantitative analysis of the characteristics identified for every lead of the ECG signal, with the objective of identifying and excluding any outliers. The characteristics taken into consideration include parameters such as the points and specific waves of the trace: T, J, R, S, P, etc. In order to perform a homogeneous comparison, the analysis is conducted on each characteristic for successive cardiac cycles, assessing the stability of their positions or values within a predefined time window. The time window used for the analysis is determined based on the duration of the data, expressed in number of heartbeats, and can be parameterised according to recording method adopted.
[0198] Within this window, for each lead, probability density functions are calculated for every identified parameter, such as, for example, the temporal position of the T wave or J point in the various cardiac cycles. The probability density functions allow for an assessment of the dispersion of the characteristics between one heartbeat and another, thus providing a quantitative measure of variability. The mean and standard deviation of probability density functions are calculated for each parameter: the mean represents the central value around which the data are distributed, whereas the standard deviation measures the extent of the dispersion, indicating how far the values of the parameters deviate from the mean. A high standard deviation of the characteristic parameters is indicative of a signal with noise or artefacts, which could compromise the diagnostic analysis. In order to establish the quality of the data sample, the percentageof points falling within an interval equal to the standard deviation above or below the mean is verified. If the percentage is less than 60% of the total, the sample will be considered unreliable and discarded, and the process will pass to the next time window for a new analysis. If, by contrast, the sample satisfies the quality criterion, it will be stored for further verifications. In this step it is possible to manually exclude the data falling outside the interval equal to two times the standard deviation above or below the mean (±2 sigma), in order to reduce the risk of including false positives or negatives due to isolated components of low quality. This approach ensures that the analysis is based on consistent, reliable data.
[0199] The second step of the quality analysis assesses the temporal coherence of the signal components between different leads. After the quality of the data has been confirmed in each lead, one proceeds by comparing consecutive heartbeats of a lead with the corresponding heartbeats in the other leads, overlaying them in time to analyse the agreement of the detected components. This overlay is done using an anchorage point as a reference, generally the R peak or S point, which acts as a common time element for aligning the different leads. The procedure is then repeated for all the ECG leads, making it possible to calculate the probability density functions for the temporal distribution of the various parameters identified. This step makes it possible to verify the accuracy of positioning of the cardiac components, since significant differences in timing between the leads can indicate measurement errors.
[0200] In addition to verifying temporal dispersion, the steps of the process relating to quality control can include a correction of the identified points, exploiting the comparison between leads to improve the precision of the results. For example, leads such as V5 and Dll can be used as main references to determine the final point of the T wave (T-endset), since they offer a clearer identification of the end of the T wave. Once the final point of the T wave has been identified in these leads, the value can be used to adapt and improve the accuracy of the parameter in the remaining leads, thereby assuring greater uniformity in the overall analysis.
[0201] After completion of the quality control step, steps of analysing a pattern in the ST segment of the electrocardiogram signal follow, as described below; it may suggest the presence of cardiac anomalies. In one example and with reference to the block diagram in figure 28B, the process can envisage the steps of:
[0202] - calculating a first derivative of the electrocardiogram signal in the ST segment (step 290),
[0203] - determining one or more critical ST points where the first derivative of the electrocardiogram signal is zero (step 291),
[0204] - analysing the sign of the first derivative of the electrocardiogram signal in each interval defined by the adjacent critical ST points (step 292).
[0205] In the event that the first derivative of the electrocardiogram signal is positive in one of said intervals comprised between adjacent critical ST points, the ST segment is increasing in that interval. On the other hand, if the first derivative of the electrocardiogram signal is negative in one of said intervals comprised between adjacent critical ST points, the ST segment is decreasing in that interval. Finally, if the first derivative of the electrocardiogram signal is zero in one of said intervals comprised between adjacent critical ST points, the ST segment is constant or horizontal in that interval.
[0206] The process can also envisage a step of geometrically analysing the electrocardiogram signal in order to provide a more in-depth assessment of the morphological characteristics of the signal itself. As shown in the block diagram in figure 29, the process can envisage determining an angle of curvature 16 (step 220), comprised between a first straightline 10 passing through the R' peak and the J point and a second straight line 11 passing through the J point and the T point, in order to assess the shape of the ST segment and the presence of any cardiac anomalies. The process can envisage calculating the angle of curvature 16 by moving anticlockwise, starting from the first straight line 10, to the second straight line 11. In one example, the process can envisage calculating the angle of curvature 16 as the sum of a first angle 12, a second angle 13 and a right angle 14, where the first and second angles are respectively a function of the inverse tangent of the changes in coordinates between the R' and J points, and a change between the J and T points. As shown in figure 30, the first angle 12 is comprised between the first straight line 10 and a horizontal straight line 15 passing through the J point, whilst the second angle can be comprised between the second straight line 22 and the abovementioned horizontal straight line 15. From a mathematical standpoint, the first and second angles 12, 13 can be calculated according to the following formulas.
[0207] For the first angle 12:
[0208] 1 = - arctan
[0209]
[0210] For the second angle 13:
[0211] 2 = 90°
[0212]
[0213] In the context of the graph in figure 30, which represents the electrocardiogram signal (ECG) with the voltage (in millivolts) on the y-axis and the time (in milliseconds) on the x-axis, the variables Xfi„
[0214]
[0215] Xy, Y), XT, and YT, indicate the coordinates of the characteristic points of the trace. In particular, the X-coordinates, and the / -coordinates specify, respectively, the temporal position and the amplitude of the electrocardiogram signal. More specifically, in the graph in figure 30:
[0216] - represent the coordinates along the X-axis and the Y-axis of the R' peak,
[0217] - Xj and Yj, indicate the coordinates along the X-axis and the Y-axis of the J point (J-point),
[0218] - XTand YT, indicate the coordinates along the X-axis and the Y-axis of the T point (T-peak).
[0219] Subsequently, the process can envisage a step 221 of evaluating whether the electrocardiogram signal conforms to an acceptability criterion if the angle of curvature is less than 200°, preferably comprised between 180° and 200°. The process can further envisage a step of communicating to the operator, for example via the screen of the electrocardiographic device 1, the angle of curvature calculated independently of the outcome of the comparison (see steps 222 in figure 29).
[0220] This comparison step enables the shape of the ST segment to be classified as concave or convex, depending on the measurement of the angle of curvature. If the angle is greater than 200°, optionally greater than 180°, the ST segment is considered concave (see figure 31), whereas an angle of less than 200°, optionally less than 180°, indicates a convex shape (see figure 32). The determination of the angle of curvature 16 and the morphology of the
[0221] electrocardiogram trace in this zone can provide the healthcare professional with relevant indications for identifying potential heart conditions.
[0222] Within the framework of the diagnosis, the process can be used, in some variant embodiments, to identify Brugada syndrome, a heart condition characterised by specific morphological anomalies in the ST segment. In such a context, the process can be configured to diagnose Brugada syndrome if the angle of curvature obtained does not conform to the established acceptability criterion, i.e. if the angle of curvature calculated is greater than 200°, optionally greater than 180°, indicating an excessive curvature of the ST segment (see step 223 in figure 29). In other words, the evaluation of the angle of curvature represents a useful element for carrying out an early and accurate diagnosis of Brugada syndrome, thus contributing to a more effective management of the patient's heart conditions.
[0223] In the context of the geometric analysis of an electrocardiogram signal (ECG), the process can further provide for the calculation of an intermediate angle 17 (step 230 of the block diagram in figure 33) to provide additional information on the morphology of the trace, especially in the segment comprised between the J point and the T point, making it possible to determine the pattern of the ST segment and to evaluate any irregularities in the signal.
[0224] In order to calculate the intermediate angle 17, the process can envisage initially identifying an intermediate point (identified in figures 34A and 34B by the reference term "middle”) along the portion of the trace comprised between the J point and the T point. The intermediate point represents a specific position within the ST segment and is used as a reference for the subsequent geometric evaluation.
[0225] Once the intermediate point has been established, the intermediate angle 17 is calculated between two straight lines: a primary straight line 18, passing through the J point and the intermediate point, and a secondary straight line 19, connecting the intermediate point to the T point. The intermediate angle 17 is thus determined by moving anticlockwise from the primary straight line to the secondary straight line, providing a measurement of the slope and relative curvature of the cardiac cycle of the signal. From a mathematical viewpoint, the calculation of the intermediate angle 17 is performed in a similar manner to what was previously described in relation to the angle of curvature 16, considering due adaptations given the geometric difference between the angle of curvature 16 and the intermediate angle 17.
[0226] Subsequently, the process can envisage a step 231 of verifying whether the intermediate angle conforms to an intermediate acceptability criterion. In particular, if the intermediate angle is less than 200°, optionally less than 180°, the intermediate angle 17 will be considered to conform to the intermediate acceptability criterion, suggesting that the morphology of the ST segment falls within the foreseen limits.
[0227] In contrast, if the intermediate angle is greater than 200°, optionally greater than 180°, the same intermediate angle will be considered not to conform to the intermediate acceptability criterion, suggesting the presence of a cardiac anomaly. The process can further envisage a step of communicating to the operator, for example via the screen of the electrocardiographic device 1, the intermediate angle calculated independently of the outcome of the comparison (see steps 232 in figure 33). In one example, if the intermediate angle 17 does not conform to the intermediate acceptability criterion, the process can be configured to diagnose a STEMI acute coronary syndrome (step 233 in figure 33).
[0228] Examples
[0229] In the present section, and with reference to figures 35 to 37, examples are shown to illustrate the previously described process of analysis for diagnosing Brugada syndrome, a right bundle branch block (RBBB) and STEMI acute coronary syndrome in a patient.
[0230] As regards the diagnosis of Brugada syndrome and with reference to the block diagram in figure 35, the process can envisage a first step of determining an R' wave (step 250) in the electrocardiogram trace. Moving on to the T wave, the process can envisage verifying that in at least one lead (except aVR) the same T wave is negative (step 251). It should be noted that the process takes into consideration that a negative T wave in the aVR lead does not constitute an anomaly, whereas in the remaining leads the presence of a negative T wave is representative of a cardiac anomaly. The process proceeds with the analysis of the QRS complex, verifying whether its duration exceeds 120 ms, if so indicating a potential disturbance in conduction (step 252). In a subsequent step 253, the process can be configured to verify whether the ST segment following the QRS complex is aligned with the isoelectric line. If it is not, the process can be configured to verify whether the J point (J-point) has a voltage value, above the isoelectric line, greater than 0.1 mV (step 254).
[0231] The process can proceed by verifying that the ST segment following the QRS complex has a descending slope (step 255). In other words, the process envisages verifying that the ST segment shows a decreasing trend over time. The process finally envisages carrying out the following steps:
[0232] - verifying whether the J point has a voltage value greater than or equal to 0.2 mV above the isoelectric line (step 256),
[0233] - verifying that the R'JT angle is convex in leads such as V1 and / or V2 (step 257),
[0234] - verifying that a lowest point of the T wave has a negative voltage value (step 258).
[0235] This sequence of verifications makes it possible to detect and automatically signal any morphological anomalies which, if present in the patient, indicate Brugada syndrome.
[0236] In a further example and with reference to the block diagram in figure 36, the process described can be used to diagnose a right bundle branch block (RBBB) in the patient in the presence of the following morphological anomalies in the electrocardiogram signal:
[0237] - determining an R' wave (step 270),
[0238] - verifying that in at least one lead (except aVR) the T wave is negative (step 271),
[0239] - verifying that the QRS complex has duration greater than 120 ms (step 272),
[0240] - verifying that the ST segment following the QRS complex is aligned with the isoelectric line (step 273).
[0241] Once the above-described criteria have been verified, the process can automatically detect and diagnose a right bundle branch block in the patient.
[0242] In a further example and with reference to the block diagram in figure 37, the process described can be used to diagnose a STEMI acute coronary syndrome if the following morphological anomalies of the electrocardiogram signal under examination occur:
[0243] - verify whether the T point (T-peak) of the T wave is positive, optionally greater than 0.1 mV, even more optionally greater than 1.5 mV (step 280),
[0244] - verify whether the ST segment following the QRS complex is aligned with the isoelectric line (step 281). If it is not, the process can be configured to carry out the following steps:
[0245] o verifying whether the amplitude of the J point (J-point) has a voltage value, above the isoelectric line, greater than 0.1 mV (step 282),
[0246] o verifying that the ST segment following the QRS complex has an ascending slope (step 283). In other words, the process envisages verifying that the ST segment shows an increasing trend over time,
[0247] o verifying whether the amplitude of the J point or the highest point of the ST segment is greater than or equal to 0.1 mV relative to the isoelectric line (step 284).
[0248] o verifying that the intermediate angle 17 (i.e. the angle subtended by a first straight line passing through the J point and the intermediate point, and a second straight line passing through the intermediate point and the T point) is convex (step 285),
[0249] o diagnosing STEMI acute coronary syndrome (step 286).
[0250] This sequence of verifications allows for automatically identifying the electrocardiogram anomalies which, if present, indicate a STEMI acute coronary syndrome.
Claims
CLAIMS1. A computer-implemented process for analysing at least one electrocardiogram trace relating to one or more leads of an electrocardiogram of a patient's heart, said process comprising the steps of:- acquiring at least one electrocardiogram signal which is variable as a function of time, said electrocardiogram signal being a sampled digital signal represented by a series of discrete points,- segmenting the electrocardiogram signal into one or more cardiac cycles, each of which representative of a single cardiac cycle,- identifying (103), for each of said one or more cardiac cycles, a QRS complex (QRS) as a segment of the electrocardiogram signal representative of ventricular depolarisation,- determining, for each of said one or more cardiac cycles, a T wave (T-wave) as a segment of the electrocardiogram signal comprised between the QRS complex (QRS) and a terminal segment of a respective one of said one or more cardiac cycles, said T wave (T-wave) being representative of ventricular repolarisation,characterised in that the process comprises a step of determining (211) a J point (J-point) of said series of discrete points of the electrocardiogram trace comprised between the QRS complex (QRS) and the T wave (T-wave) at which a change of concavity occurs in the same electrocardiogram trace.
2. The process according to the preceding claim, wherein the step of determining the J point (J-point) comprises the sub-steps of:- determining, among said series of discrete points, an S point (S-peak) of the QRS complex (QRS), representative of depolarisation of basal portions of the patient's heart,- determining, among said series of discrete points, a T point (T-peak) of the T wave (T-wave) representative of ventricular repolarisation of the patient's heart,- determining a J point search segment of the electrocardiogram trace comprised between the S point (S-peak) and the T point (T-peak),- calculating, among said series of discrete points, one or more inflection points of the J point (J-point) search segment,- identifying the J point (J-point) in one of said one or more inflection points.
3. The process according to the preceding claim, wherein calculating said one or more inflection points of the J point (J-point) search segment comprises:- calculating a second derivative of the J point (J-point) search segment of the electrocardiogram signal,- calculating one or more noteworthy J points (J-point) wherein the second derivative of the J point search segment is zero, does not exist or is not defined,- identifying, among said one or more noteworthy J points (J-point), an inflection point if a change of sign of the second derivative of the J point search segment occurs in said one or more noteworthy J points.
4. The process according to one of the two preceding claims, comprising a step of determining the presence or absence of an R' peak (R'-peak) representative of a second ventricular depolarisation, said R' peak being located immediately following the S point (S-peak) and within the QRS complex,wherein the electrocardiogram signal has a first derivative equal to zero at the R' peak (R'-peak) and has, in a point immediately following the R' peak (R'-peak), a negative first derivative,wherein if the step of determining the R' peak (R'-peak) gives a positive outcome, the J point (J-point) search segment of the electrocardiogram trace is comprised between the R' peak and the T point (T-peak),wherein if the step of calculating one or more inflection points of the J point search segment results in the identification of a single inflection point, then the step of identifying the J point envisages marking said single inflection point as the J point (J-point), orif the step of calculating one or more inflection points of the J point (J-point) search segment results in the identification of a plurality of inflection points, then the step of identifying the J point (J-point) envisages marking as the J point (J-point):- the point with the first derivative of the J point search segment equal to zero, or- the point, among said discrete points, with a maximum curvature comprised between the R' peak or the S point, and the T point (T-peak).
5. The process according to any one of the preceding claims when dependent on claim 2, wherein the T wave(T-wave) is comprised between an initial point of the T wave (T-onset) of said series of discrete points and a final point of the T wave (T-endset) of said series of discrete points, wherein determining the T wave (T-wave) comprises the steps of:- marking, as the final point of the T wave (T-endset), a stationary T point of said series of discrete points, wherein a first derivative of the segment of the electrocardiogram signal comprised between the T point (T-peak) and a terminal segment of a respective cycle of said one or more cardiac cycles is zero,- marking, as the initial point of the T wave (T-onset), a point of said series of discrete points defining a JT segment (JT-wave) of the electrocardiogram trace comprised between the J point (J-point) and the T point (T-peak), wherein a change of concavity occurs in the same JT segment (JT-wave),wherein marking the initial point of the T wave (T-onset) comprises:- calculating a second derivative of the JT segment (JT-wave),- analysing the sign of the second derivative of the JT segment (JT-wave) at the T point (T-peak) in order to determine a concavity of the cardiac cycle at the T point,- calculating one or more critical T points of said series of discrete points, wherein the second derivative of the JT segment (JT-wave) is equal to zero, does not exist or is not defined,- analysing the sign of the second derivative of the JT segment (JT-wave) at each of said one or more critical T points to determine a concavity of the cardiac cycle in said one or more critical T points,identifying, between said one or more critical T points, the initial point of the T wave (T-onset) as the critical T point where the curvature of the JT segment (JT-wave) has maximum absolute value and concavity with a sign opposite the one calculated at the T point (T-peak).
6. The process according to any one of the preceding claims when dependent on claim 2, wherein determining the T point envisages marking the T point (T-peak) as an absolute maximum or minimum point of the T wave (T-wave), wherein determining the T point (T-peak) comprises a sub-step of calculating, among said series of discrete points, one or more stationary T points wherein a first derivative of the segment of the electrocardiogram signal comprised between the QRS complex (QRS) and a terminal segment of a respective one of said one or more cardiac cycles is zero,whereinif the step of calculating said one or more stationary T points results in the identification of a single stationary point, then the process envisages marking said single stationary point as the T point (T-peak), orif the step of calculating said one or more stationary T points results in the identification of a plurality of stationary points, then the process comprises:- determining a stationary T point with a maximum value and a stationary T point with a minimum value among the plurality of stationary T points,- marking as the T point (T-peak) the stationary T point with the minimum value if the stationary T point with the maximum value is negative relative to an isoelectric line of the electrocardiogram signal representative of non- detectable cardiac action potentials,- marking as the T point (T-peak) the stationary T point having the maximum value if the stationary T point with the minimum value is positive relative to the isoelectric line,- marking as the T point (T-peak) the higher in absolute value between the stationary T point with the minimum value and the stationary T point with the maximum value if the stationary T point with the maximum value is positive relative to the isoelectric line and the stationary T point with the minimum value is negative relative to the isoelectric line.
7. The process according to any one of the preceding claims, comprising a procedure for calculating the negative curvature of the cardiac cycle of the electrocardiogram signal, said procedure for calculating the negative curvature having the steps of:- calculating, for every point of said series of discrete points below the isoelectric line, a curvature of the entire electrocardiogram signal,- calculating a negative curvature threshold value as the arithmetic mean of said curvatures of the entire electrocardiogram signal,wherein the process further comprises a procedure for calculating the positive curvature having the steps of:- calculating, for every point of said series of discrete points above the isoelectric line, a curvature of the entire electrocardiogram signal,- calculating a positive curvature threshold value as the arithmetic mean of said curvatures of the entire electrocardiogram signal,wherein when the present claim depends on claim 2, determining an S point of the QRS complex (QRS) comprises the steps of:- determining one or more noteworthy S points among said series of discrete points of the cardiac cycle, where the first derivative of the cardiac cycle of the electrocardiogram signal is zero,- comparing, for each of said one or more noteworthy S points, the curvature of the cardiac cycle in said one or more noteworthy S points with the negative curvature threshold value,- identifying one or more minimum points between said one or more noteworthy S points in which the cardiac cycle has a curvature value greater than the negative curvature threshold value and / or having a negative voltage value.
8. The process according to the preceding claim, wherein the process comprises a step of determining a highest positive peak of the cardiac cycle of the electrocardiogram signal representative of an absolute maximum point of said cardiac cycle,wherein determining an S point (S-peak) of the QRS complex (QRS) further comprises the steps of:- marking the S point (S-peak) as overlying the highest positive peak of the cardiac cycle of the electrocardiogram signal if the cardiac cycle, in none of said one or more noteworthy S points, has a curvature value greater than the negative curvature threshold value, or- marking as the S point (S-peak) the point between said one or more minimum points, adjacent to the highest positive peak.
9. The process according to the preceding claim, wherein if identifying one or more minimum points between said one or more noteworthy S points determines the presence of a plurality of noteworthy S points downstream of the highest positive peak, determining an S point further comprises the steps of:- verifying that between the S point (S-peak) and the highest positive peak there is no presence of a further positive peak, representative of a relative maximum point of the cardiac cycle of the electrocardiogram signal with a curvature greater than the positive curvature threshold value,- if the step of verifying that between the S point (S-peak) and the highest positive peak gives a negative outcome, the process envisages marking as the S point (S-peak) a further of said one or more minimum S points and subsequently carrying out the steps of:o reiterating the step of verifying that no further positive peak is present between the S point (S-peak) and the highest positive peak, and / oro marking the S point (S-peak) as overlying the highest positive peak in the event that there are no further minimum points,wherein the step of verifying that no further positive peak is present between the S point (S-peak) and the highest positive peak envisages, as the first step, verifying the most negative peak between said one or more minimum points of said one or more noteworthy S points.
10. The process according to one of the two preceding claims, wherein if identifying one or more minimum points between said one or more noteworthy S points determines the presence of a plurality of noteworthy S points upstream of the highest positive peak, determining an S point further comprises the steps of:- verifying the presence or absence of a further positive peak, representative of a relative maximum point of the cardiac cycle of the electrocardiogram signal with a curvature greater than the positive curvature threshold value, located upstream of the S point (S-peak),wherein if the step of verifying the presence or absence of a further positive peak upstream of the S point (S-peak) gives a negative outcome, the process envisages marking as the S point (S-peak) a further of said one or more minimum S points and subsequently carrying out the steps of:- reiterating the step of verifying the presence or absence of a further positive peak upstream of the S point (S-peak), and / or- marking the S point (S-peak) as overlying the highest positive peak in the event that there are no further maximum points.
11. The process according to any one of the three preceding claims, wherein if identifying one or more minimum points between said one or more noteworthy S points determines a first plurality of minimum points comprised between the highest positive peak and the terminal segment of a respective one of said one or more cardiac cycles and a second plurality of minimum points comprised between the highest positive peak and the initial segment of a respective one of said one or more cardiac cycles, determining an S point further comprises the steps of:- selecting the most negative point among said first plurality of minimum points,- selecting the most negative point among said second plurality of minimum points,- determining one or more positive QRS peaks representative of relative maximum points of a respective one of said one or more cardiac cycles of the electrocardiogram signal,- calculating a curvature of the cardiac cycle of the electrocardiogram signal at said one or more positive QRS peaks,- comparing the curvature of said one or more positive QRS peaks with the positive curvature threshold value, wherein determining the S point further comprises the steps of:- if each of said one or more positive QRS peaks is less than the positive curvature threshold value, marking as the S point (S-peak) the most negative point of said one or more minimum points of the first plurality, or- if the step of determining one or more positive QRS peaks results in the identification of a single positive QRS peak, optionally a single relative maximum point, marking as the S point (S-peak) the minimum S point comprised between said positive QRS peak, optionally said relative maximum point, and the highest positive peak.
12. The process according to any one of the preceding claims when dependent on claim 14, comprising a step of determining an R peak (R-peak) as a maximum point of the QRS complex (QRS),optionally wherein the R peak (R-peak) is comprised between an initial point of the QRS complex (QRS) of said series of discrete points and the S point (S-peak),wherein determining an R peak (R-peak) comprises the sub-steps of:- calculating a first derivative of the complex (QRS),- determining, among said series of discrete points, one or more critical R points where the first derivative of the complex (QRS) is zero, wherein said one or more critical R points are above a / the isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials,- calculating a curvature of the complex (QRS) in said one or more critical R points,- comparing the curvature of the complex (QRS) in said one or more critical R points with the positive curvature threshold value,wherein determining an R peak (R-peak) further comprises the sub-steps of:- marking as the R peak (R-peak) the highest of said one or more critical R points where the curvature of the complex (QRS) is greater than the positive curvature threshold value, or- marking as the R peak (R-peak) a point between said one or more critical R points in which the QRS complex (QRS) has zero curvature.
13. The process according to any one of the preceding claims when dependent on claim 4, wherein determining an R' peak (R'-peak) comprises the sub-steps of:- calculating a first derivative of a segment of the cardiac cycle comprised between the S point (S-peak) and an endpoint of the complex (QRS),- determining, among said series of discrete points, one or more critical R' peaks where the first derivative of the complex (QRS) is zero, wherein said one or more critical R' peaks are above a / the isoelectric line of the electrocardiogram signal representative of non-detectable cardiac action potentials,- analysing the sign of the first derivative of a segment of the cardiac cycle comprised between:o the critical R' peak having the maximum value and the endpoint of the complex (QRS), oro the critical R' peak having the maximum value and the point between said one or more discrete points distant from said critical R' peak by a time interval comprised between 0.01s and 0.1s,wherein determining an R' peak (R'-peak) further comprises the sub-steps of:- marking as the R' peak (R'-peak) the critical R' peak having the maximum value if the sign of the first derivative of the segment of the cardiac cycle comprised between the R' peak with the maximum value and the endpoint of the complex (QRS) is negative, or- marking the R' peak (R'-peak) as coinciding with the S point (S-peak).
14. The process according to any one of the preceding claims, comprising the step of:- calculating an angle of curvature comprised between a first straight line passing through the R' peak (R'-peak) and the J point (J-point) and a second straight line passing through the J point (J-point) and the T point (T-peak), optionally determining whether the angle of curvature conforms to an acceptability criterion if the same angle of curvature is less than 200°, optionally comprised between 180° and 200°.wherein the angle of curvature is calculated proceeding anticlockwise starting from the first straight line towards the second straight line.
15. The process according to any one of the preceding claims comprising the steps of:- determining an intermediate point of the cardiac cycle of the electrocardiogram signal between the J point (J-point) and the T point (T-peak),- calculating an intermediate angle comprised between a primary straight line passing through the J point (J-point) and the intermediate point and a secondary straight line passing through the intermediate point and the T point (T- peak),optionally determining whether the intermediate angle conforms to an intermediate acceptability criterion if the intermediate angle is less than 180°,wherein the intermediate angle is calculated proceeding anticlockwise starting from the primary straight line towards the secondary straight line.
