WO2017182622A1

WO2017182622A1 - Adaptive visualization of electrocardiogram

Info

Publication number: WO2017182622A1
Application number: PCT/EP2017/059507
Authority: WO
Inventors: Richard Earl GREGG
Original assignee: Koninklijke Philips N.V.
Priority date: 2016-04-21
Filing date: 2017-04-21
Publication date: 2017-10-26

Abstract

A method for digital visualization of electrocardiography (ECG) data receives ECG data of electrical activity of a heart of a patient collected over a period of time. The ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes. The method calculates one or more ECG statistics of the ECG data and determines an ECG interpretation using the ECG statistics. The method also selects, using a mapping between a set of ECG interpretations and a set of report formats, a report format corresponding to the ECG interpretation, converts the ECG data according to the report format to produce an ECG report, and renders the ECG report to an output interface.

Description

Adaptive visualization of electrocardiogram

Technical Field

[0001] This disclosure concerns electrocardiography. More particularly, but not exclusively, the present disclosure concerns digital visualization of electrocardiograms.

Background

[0002] e.g. Twelve-lead electrocardiogram (ECG) reports are typically displayed on a grid of four columns by three rows, with each square showing a 2.5 second tracing for each of twelve signals: three limb leads (I, II, III), three augmented limb leads (aVR, aVL, and aVF), and six precordial leads (V1-V6). While this 12-lead display is often the default format, it may be either not useful or less than ideal for assessing some patient conditions or situations. Different abnormalities of the heart can be determined (or more easily assessed by a clinician) from different combinations of the ECG tracings of different durations. For example, the complete 10 sec display of the ECG tracing from a lead VI may allow more confident diagnosis of atrial fibrillation. To that end, the ECG tracings of other leads may be unnecessary or less important to display in the face of limited interface real estate. However, the ECG tracing of the lead VI may be less useful to detect the premature ventricular beats, but can be (more easily) detected from ECG tracings of other leads. SUMMARY

[0008] Some embodiments are based on recognition that a single electrocardiogram (ECG) and/or an ECG report can be suboptimal for showing all different variations of cardiovascular abnormalities. This problem is rooted in low dimensionality of an interface for digital visualization of electrocardiograms as contrasted with dimensions of the ECG data. For example, the two-dimensional (2D) computer screen can be suboptimal in displaying all intricacies of 12D ECG data from 12- lead ECG system. To that end, some embodiments are based on understanding that different cardiovascular abnormalities can be better seen on different formats of ECG reports.

[0009] For example, the atrial signal from atrial fibrillation is may be better visualized in an ECG tracing from a lead VI, while the ECG tracings from the leads II, III and aVF may be better for recognizing atrial flutter. Similarly, premature ventricular beats (PVC) can be seen in the ECG tracings from the chest leads, while the ECG tracings from all twelve leads can be viewed simultaneously at high gain to detect the earliest Q- or R-wave and the lead having the ECG tracing with the longest end of T-wave to measure QT interval.

[0010] However, the mere availability of different formats of ECG reports for detecting different abnormalities does not guarantee the usage of the correct format of the ECG report that is better suited to detect the particular cardiovascular abnormality. For example, some ECG readers do not know all of the features of the ECG reading/viewing programs. For example, although alternate ECG report formats may be available, an ECG reader can use a standard ECG report that includes the ECG tracings from all twelve leads taken for a standard period of time, e.g., 10 sec. Such a situation leads to information overflow problem, such as generating the ECG report that includes information not needed for medical diagnostic, which consumes unnecessary computational and memory resources of a digital visualization system.

[0011] Also, if the ECG reader does know about a particular format of the ECG report, it may takes too long and too many mouse clicks (or similar computer operations or other effort on the part of the reader) to make a particular ECG report format appear. The ECG readers typically read ECG reports very fast and may not take the time to switch between different formats of the ECG report. However, the usage of the suboptimal format of ECG report can lead to misinterpretation of the ECG rhythm and morphology.

[0012] According to some embodiments, systems may automatically select an appropriate format of ECG report based on conditions or features automatically identified in the ECG data. For example, an automated program can interpret the ECG data to generate an "ECG interpretation" (e.g., "sinus rhythm," "Atrial fibrillation," etc.) and to automatically select an ECG report format suitable for such an ECG interpretation. For example, a report format that corresponds to the ECG interpretation can be selected from a mapping between a set of ECG interpretations and a set of report formats. For example, such a mapping can be implemented as a lookup table that correlates the ECG interpretation to a report format (e.g., "10 sec of leads V1-V6" or "12x1"). The ECG data can be converted to an ECG report according to the selected report format and rendered to the ECG reader. In such a manner, the ECG reader is automatically presented with the ECG report suitable, e.g.., optimized, for the particular cardiovascular abnormality, which reduces the computation and memory usage of the digital visualization system and can reduce the misinterpretation of the ECG data. In some situations, even if the automated analysis is incorrect, the optimal ECG report for that incorrect abnormality can allow the ECG reader to promptly recognize the incorrect interpretation.

[0013] For example, the ECG interpretation can be determined using the ECG statistics of the ECG data, such as one or combination of a baseline content, an average amplitude, a peak amplitude, an average frequency, a R-wave count, and an ECG rate of the ECG data. For example, the ECG statistics can be analyzed using one or combination of a probabilistic Bayesian analysis, cluster analysis, artificial neural networks, and regression analysis to determine the ECG interpretation. Additionally or alternatively, one embodiment determines an ECG interpretation using a set of expert rules where each rule is a set of criteria where individual calculated ECG features such as heart rate and R- wave amplitude are tested against thresholds and combined with logical "and", "or" and "not." [0014] In some embodiments, the report format includes a type of the report and a layout of the report. To that end, those embodiments can select a subset and duration of ECG tracings according to the layout of the report and convert the subset of ECG tracings into a set of mutually arranged images according to the type of the report. In such a manner, the memory requirements for storing multiple ECG formats are further reduced.

[0015] Accordingly, various embodiments disclosed herein are directed to a method for digital visualization of electrocardiography (ECG) data that includes receiving ECG data of electrical activity of a heart of a patient collected over a period of time, wherein the ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes; calculating one or more ECG statistics of the ECG data; determining an ECG interpretation using the ECG statistics; selecting, using a mapping between a set of ECG interpretations and a set of report formats, a report format corresponding to the ECG interpretation; converting the ECG data according to the report format to produce an ECG report; and rendering the ECG report to an output interface. The steps of the method are performed by a processor.

[0016] Various embodiments disclosed herein are directed to a system for digital visualization of electrocardiography (ECG) data that includes a memory to store a mapping between a set of ECG interpretations and a set of report formats; an output interface to render an ECG report; a network interface to receive ECG data of electrical activity of a heart of a patient collected over a period of time, wherein the ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes; and a processor configured to calculate one or more ECG statistics of the ECG data; determine an ECG interpretation using the ECG statistics; select, using the mapping, a report format corresponding to the ECG interpretation; convert the ECG data according to the report format to produce the ECG report; and render the ECG report to the output interface.

[0017] Various embodiments described herein are directed to a non-transitory computer readable storage media embodied thereon a program executable by a processor for performing a method that includes receiving ECG data of electrical activity of a heart of a patient collected over a period of time, wherein the ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes; calculating one or more ECG statistics of the ECG data; determining an ECG interpretation using the ECG statistics; selecting, using a mapping between a set of ECG interpretations and a set of report formats, a report format corresponding to the ECG interpretation; converting the ECG data according to the report format to produce an ECG report; and rendering the ECG report to an output interface. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A is a block diagram of a computer system for digital visualization of electrocardiography (ECG) data in accordance with some embodiments.

[0019] Figure IB is a schematic of a medical system according to one embodiment.

[0020] Figure 2 is a block diagram of a method for digital visualization of the ECG data according to some embodiments.

[0021] Figure 3 is an example of implementation of the mapping used by the computer system of Figure 1 A according to one embodiment.

[0022] Figure 4A is a schematic of training a regression function according to one embodiment.

[0023] Figure 4B is a schematic of the training a neural network used by some embodiments.

[0024] Figure 5 is a flow chart of a method for automatically determining an ECG interpretation according to some embodiments.

[0025] Figure 6 is a block diagram of a method for updating the mapping according to some embodiments.

[0026] Figures 7A and 7B are examples of the ECG reports used for digital visualization of atrial fibrillation and atrial flutter according to some embodiments.

[0027] Figures 8A and 8B are different examples of the ECG reports according to some embodiments. [0028] Figures 9A and 9B are examples of the ECG reports used for digital visualization of QT interval according to some embodiments.

[0029] Figures 10A, 10B and IOC are examples of the ECG reports used for digital visualization of T-wave abnormality according to some embodiments.

[0030] Figures 11A, 11B and 11C are different examples of QRS and T-wave loops for a normal adult ECG.

[0031] Figures 12A and 12B are examples of ECG reports used for digital visualization of acute coronary syndrome according to some embodiments.

DETAILED DESCRIPTION

[0032] ECG reports used by electrocardiographers to read ECGs provide a compromise between the gain, the duration and the number of the ECG tracings. Such ECG reports may show long enough segments of the ECG tracings that include multiple beats, which are necessary to confirm that what is seen in one beat is seen in the other beats and the abnormality is a true physiological feature and not just a random short term artefact. However, the standard ECG reports do not work for some cases. For example, in some situations, the ECG rhythm and morphology in the standard reports can be misinterpreted because the view of the signals is not ideal for that particular situation. Conversely, in some other situations, the standard ECG reports can provide unnecessary information. However, generating the ECG report that includes information not needed for medical diagnostic consumes unnecessary computational and memory resources. In view of the foregoing, it would be desirable to adapt ECG reports to the multitude of different cardiovascular abnormalities.

[0033] Figure 1A shows a block diagram of a computer system 100 for digital visualization of electrocardiography (ECG) data in accordance with some embodiments. The computer system 100 includes a processor 102 configured to execute stored instructions, as well as a memory 104 that stores instructions that are executable by the processor. As used herein, the term processor will be understood to encompass a single core microprocessor, a multi-core microprocessor, a computing cluster, a field programmable gate array (FPGA), application specific integrated circuit (ASIC), or any number of other configurations including combinations thereof. The memory 104 can include random access memory (RAM), read only memory (ROM), flash memory, or any other suitable memory (including storage) systems. As used herein, the term "non- transitory computer-readable storage medium" will be understood to encompass volatile memory (e.g., DRAM and SRAM) and non-volatile memory (e.g., flash, magnetic, and optical memories) but to exlude transitory signals. The processor 102 is connected through a bus 106 to one or more input and output devices.

[0034] In some embodiments, the memory 104 is a non- transitory computer readable medium embodied thereon a digital visualization program executable by a processor, such as the processor 102, for performing a method for digital visualization of the ECG data. The computer system 100 can also include a storage device 108 operatively connected to the memory 104. The storage device 108 can include a hard drive, an optical drive, a thumbdrive, an array of drives, or any combinations thereof. In some implementations, the digital visualization program and/or other data needed for the execution of the digital visualization program are stored on one or combination of the memory 104 and the storage device 108 collectively referred herein as a memory.

[0035] In some embodiments, the method for digital visualization of the ECG data automatically selects an appropriate format of ECG report based on conditions or features automatically identified in the ECG data. For example, the digital visualization program can include an automated ECG interpretation program 110 that can interpret the ECG data 138 to generate an "ECG interpretation" (e.g., "sinus rhythm," "Atrial fibrillation," etc.). For example, a format of an ECG report that corresponds to the ECG interpretation can be selected from a mapping 112 between a set of ECG interpretations and a set of report formats. For example, such a mapping can be implemented as a lookup table that correlates the ECG interpretation (e.g., a token, enumerated value, string, or other value output by the ECG interpretation program 110) to a report format (e.g., "10 sec of leads V1-V6" or "12x1"). The ECG data can be converted, using ECG conversion program 114, to an ECG report according to the selected format of the ECG report. In such a manner, an ECG report reader may be automatically presented with the ECG report optimized for the particular cardiovascular abnormality (as identified by the ECG interpretation program 110), which reduces the computation and memory usage of the digital visualization system and can reduce the misinterpretation of the ECG data.

[0036] The computer system 100 can also include a human machine interface 116 within to connect the system to, for example, a keyboard 118 and pointing device 120, wherein the pointing device 120 can include a mouse, trackball, touchpad, joy stick, pointing stick, stylus, or touchscreen, among others. The computer system 100 can be linked through the bus 106 to a one or different output interfaces for rendering the ECG report to the ECG report reader. For example, the output interface can include a display interface 122 adapted to connect the system 100 to a display device 124. The display device 124 can include a computer monitor, camera, television, projector, or mobile device, among others.

[0037] Additionally or alternatively, the output interface can include an imaging interface 126 adapted to connect the system to an imaging device 128. The imaging device 128 can include a camera, computer, scanner, mobile device, webcam, projector, or any combination thereof. Additionally or alternatively, the output interface can include a printer interface 130 connected to the computer system 100 through the bus 106 and adapted to connect the computer system 100 to a printing device 132. The printing device 132 can include a liquid inkjet printer, solid ink printer, large-scale commercial printer, thermal printer, UV printer, or dye-sublimation printer, among others. [0038] A network interface 134 is adapted to connect the computer system 100 through the bus 106 to a network 136. Through a communication link formed by the network 136, the ECG data 138 including one or combination of the ECG tracings can be downloaded and stored within the computer's storage system 108 for storage, processing and visual representation.

[0039] In some embodiments, the network interface 134 can establish a communication link with a medical device including a set of electrodes configured to detect ECG signals of the heart of the patient. The communication link can establish wired or wireless connection with the medical device through, e.g., intranet or the Internet. In some embodiments, the medical device is integrated with the system 100, such that the communication link is implemented through the bus 106.

[0040] Figure IB shows a schematic of a medical system 20 according to one embodiment. The medical system includes a medical device, such as an ECG monitor 28, for measuring the ECG of a patient 24. The ECG monitor 28 uses one or more electrodes 32 attached to the patient's body. The electrodes sense the electrical activity of the patient's heart and produce corresponding electrical signals, referred to herein as ECG signals. The ECG signals are provided to the ECG monitor via a cable 36. The ECG monitor typically outputs ECG traces that plot the ECG signals as a function of time. The ECG monitor can include an analog-to-digital converter to covert the ECG signals into the ECG data 138. An ECG reader 42, typically a cardiologist or other physician, examines the ECG signals and attempts to identify cardiac conditions, such as cardiac events or pathologies, which are of interest.

[0041] Figure 2 shows a block diagram of a method for digital visualization of the

ECG data according to some embodiments. The steps of the method can be performed by a processor, e.g., the processor 102 of the system 100. The method receives, e.g., over the communication link, ECG data 138 of electrical activity of a heart of a patient collected over a period of time. The ECG data include digital representation of a set of ECG tracings of electrical activity of a heart of a patient.

[0042] The method calculates 220 one or more ECG statistics 225 of the ECG data.

For example, the ECG statistics 225 can include one or combination of a baseline content, an average amplitude, a peak amplitude, an average frequency, an R-wave count, and a heart rate of the ECG data. The method determines 230 an ECG interpretation 235 using the ECG statistics 225. For example, the method can determine the ECG interpretation 235 using one or combination of a deterministic analysis of expert rules or a probabilistic Bayesian analysis, cluster analysis, artificial neural networks, and regression analysis.

[0043] The method selects 240 a report format 245 corresponding to the ECG interpretation 235. In various embodiments, the selection 240 uses a mapping 112 between a set of ECG interpretations and a set of report formats. Next, the method converts 250 the ECG data 138 according to the report format 245 to produce an ECG report 260 and renders the ECG report 260 to an output interface, such as the interfaces 122, 126 or 130. In some embodiments, the ECG report 260 is a digital image converted from at least part of the ECG data.

[0044] Figure 3 shows an example of implementation of the mapping 112 according to one embodiment. In this embodiment, the mapping 112 is implemented as a lookup table 300 that correlates the ECG interpretation 310 to a report format 340. In one embodiment, the report format includes a type 320 of the report, e.g., a rhythm, morphology, and a VCG type, and a layout 330 of the report, e.g., "10 sec of leads V1-V6" or "12x1". In some embodiments, the type of the report can indicate the subset of the ECG tracings and the duration of those tracings, i.e., a period of time for which the ECG tracings are recorded. The layout of the report can indicate format and mutual arrangements of the selected ECG tracings. The selection of the subset of ECG tracings can further reduce the memory requirements of the system 100.

[0045] One embodiment determines the ECG statistic 225 including one or combination of baseline content, average amplitude, average frequency, R-wave count, and ECG rate, which are next described in more detail.

[0046] The baseline content of an ECG is a measure of the ratio of low-slope ECG data to high-slope ECG data. Low-slope ECG data is generally indicative of a flat baseline in the patient's ECG while high-slope ECG data is generally indicative of a QRS complex. Low-slope and high-slope ECG data may be determined by first dividing the segment of ECG data into smaller portions (e.g., at points within the ECG where the slope of the ECG data changes from positive to negative or vice versa). The slope of each of these smaller portions is then calculated. The number of portions having low slopes compared to high slopes is counted. For example, in one embodiment, the portions having a slope less than about 1/16 of the highest slopes are counted as a value Histol, while the portions having a slope less than about 1/8 of the highest slopes are counted as a value Histo2. If either Histol or Histo2 exceed specified limits, the system 100 considers the ECG data as having high baseline content.

[0047] The average amplitude, as the name suggests, is a measure of the average amplitude of the ECG data. The average amplitude may be calculated by summing absolute values of the data points in the segment of ECG data, and dividing the result by the number of data points. Alternatively, intermediate average amplitude values may be calculated for each of the ECG waves or smaller portions of the ECG segment (defined above for calculating baseline content) with an overall average amplitude calculated from the intermediate average amplitude values. The average amplitude measure is preferably scaled to a predefined range of values.

[0048] The average frequency, as the name suggests, is a measure of the average frequency of the ECG. The average frequency may be calculated by dividing a weighted summation of the derivative of the ECG data by a weighted summation of the absolute values of the ECG data. The average frequency measure is also preferably scaled to a predefined range of values. [0049] The R-wave count is a measure of the number of QRS complexes that appear to be present in the segment of ECG data. The positive and negative-sloped portions of the ECG segment (previously defined with respect to calculating baseline content) may be evaluated separately to produce two R-wave counts, namely, PRNUM (for positive-sloped portions) and NRNUM (for negative-sloped portions). An ECG portion is designated as indicating a QRS complex (thus incrementing the R-wave count) when the average slope of the portion exceeds a predefined limit.

[0050] The heart rate is a measure that estimates the rate of heart beats occurring in the ECG data. Not all of the portions of the ECG data previously defined with respect to calculation of baseline content need to be used to calculate the ECG rate. For instance, the heart rate may be calculated using only the portions of the ECG data having a peak- to-peak amplitude and peak slope that exceed predetermined minimum peak-to-peak amplitude and slope thresholds. One actual embodiment of the defibrillator 10 uses the portions of the ECG data that are in the upper 75% of peak-to-peak amplitude and in the upper 67% of peak slope to calculate ECG rate. In addition, the positive and negative sloped portions may be analyzed separately, producing two estimates of the ECG rate. In that regard, the lower of the two estimates is preferably used as the ECG rate measure.

[0051] Some embodiments process the ECG statistic 225 to determine 230 the ECG interpretation 235 using automated ECG interpretations. Automated ECG interpretation may include use of artificial intelligence and pattern recognition software and knowledge bases to carry out automatic interpretation, test reporting, and computer-aided diagnosis of electrocardiogram tracings obtained usually from a patient. For example, some embodiments use one or combination of a probabilistic Bayesian analysis, cluster analysis, artificial neural networks, and regression analysis.

[0052] Figure 4A shows a schematic of training 401 the regression function 411 according to at least one embodiment. The regression function establishes a correspondence 405 between one or more of the ECG statistics 415 and the ECG interpretations 416. Knowing the regression function 411, the particular ECG interpretation 421 can be determined from the particular combination of the ECG statistics 430. The regression function 411 can be any complex function. For example, the regression function can be linear, nonlinear, and nonparametric regression function. In some embodiments, the regression function can be a polynomial function or a spline.

[0053] Figure 4B shows a schematic of the training of a neural network used by some embodiments of the invention. The training 451 uses a correspondence 405 between one or more of the ECG statistics 415 and the ECG interpretations 416 to produce the weights 460 of the neural network. In general, training an artificial-neural-network comprises applying a training algorithm, sometimes referred to as a "learning" algorithm, to an artificial-neural-network in view of a training set. A training set may include one or more sets of inputs and one or more sets of outputs with each set of inputs corresponding to a set of outputs. A set of outputs in a training set comprises a set of outputs that are desired for the artificial-neural-network to generate when the corresponding set of inputs is inputted to the artificial -neural-network and the artificial- neural-network is then operated in a feed-forward manner. Training the neural network involves computing the weight values associated with the connections in the artificial- neural-network. To that end, unless herein stated otherwise, the training includes electronically computing weight values for the connections in the fully connected network, the interpolation and the convolution.

[0054] Figure 5 shows a flow chart of a general method for automatically determining the ECG interpretation according to some embodiments. A digital representation of each recorded ECG channel is obtained 510, e.g., using an analog-digital convertor and/or a digital signal processing (DSP) chip. The resulting digital ECG data is processed by a series of specialized filters 520 to, e.g., reduce noise and the baseline wander of the ECG data. Different ECG features of the ECG data are extracted 530 from the filtered ECG data to identify and measure a number of ECG statistics used by various embodiments, such as the peak amplitude, area under the curve, displacement in relation to baseline, etc., of the P, Q, R, S and T waves, the time delay between these peaks and valleys, heart rate frequency (instantaneous and average), and many others. A secondary processing such as Fourier analysis and wavelet analysis may also be performed on the ECG statistics in order to provide input to pattern recognition-based programs 540. [0055] Various embodiments perform 540 logical processing and pattern recognition using, e.g., rule-based expert systems, probabilistic Bayesian analysis, fuzzy logics analysis, cluster analysis, artificial neural networks, and others methods to derive conclusions, interpretation and diagnosis of the ECG data to form 550 an automated ECG interpretation 560.

[0056] Some embodiments are based on recognition that the automated ECG interpretation, as described for example in relation with Figures 4A, 4B and 5, allows determining custom report layouts specifically tailored for a particular ECG interpretation. For example, the regression function 411 and/or the weights 460 can be trained and/or retrained for a set of different ECG interpretations and customs report formats can be created for one or a subset of the ECG interpretation. To that end, some embodiments can update the mapping 112 and/or the lookup table 300 with the new report format for the existing ECG interpretation or with new or existing report formats for the new ECG interpretation.

[0057] Figure 6 shows a block diagram of a method for updating the mapping 112 according to some embodiments. The method determines 610 a custom ECG report 615 tailored, i.e., determined 620 for a particular ECG interpretation 625, and updates 630 the mapping with a correspondence between the particular ECG interpretation and the custom report layout. For example, if the mapping 112 is implemented as a lookup table 300, the update 630 adds an additional record to the lookup table corresponding to the ECG interpretation 625 and the RCG report 615. For example, if the mapping 112 is implemented as a mapping function, the update 630 modifies the function. Figures below show some non-limited examples of custom ECG reports used by various embodiments.

[0058] Figures 7 A and 7B show examples of the ECG reports for atrial fibrillation and atrial flutter according to some embodiments. For example, Figure 7 A show an ECG report having a column grid layout and Figure 7B shows an ECG report having a custom rhythm layout. The atrial signal from atrial fibrillation is often visualized in lead VI while leads II, III and aVF are often used for recognizing the atrial flutter. Because the atrial fibrillation and atrial flutter are sometimes difficult to discriminate, the custom rhythm report of Figure 7B includes ECG tracings from leads which are advantageous for diagnosing both atrial flutter and atrial fibrillation. In this example, the complete 10 sec display of lead VI allows confident diagnosis of atrial fibrillation. The arrows 710 under the ECG tracing for the lead VI indicate the clear cyclic atrial activity.

[0059] Figures 8A and 8B show different examples of the ECG reports according to some embodiments. The ECG tracings received from the chest leads VI to V3 and V4 to V6 cannot be seen until time 5.0 sec and 7.5 sec respectively in the 3 row 4 column grid layout of the ECG report of Figure 8A. For example, premature ventricular beats (PVC) that happen before time 5 or 7.5 sec cannot be seen in those leads. In this example, the PVC can only be seen in the chest leads and looks normal in the leads that are visualized early in the ECG report of Figure 8A. To that end, for the purposes of analyzing the PVC, the report of Figure 8A can mislead the ECG reader because the necessary leads are not seen in this report. However, the rhythm layout of the ECG report of Figure 8B shows that the first beat is a PVC.

[0060] Figures 9A and 9B show examples of the ECG report for analyzing QT interval according to some embodiments. In cardiology, the QT interval is a measure of the time between the start of the Q wave and the end of the T wave in the heart's electrical cycle. The QT interval represents electrical depolarization and repolarization of the ventricles. A lengthened QT interval is a marker for the potential of ventricular tachyarrhythmias like torsades de pointes and a risk factor for sudden death. In some embodiments, to manually measure QT interval with the acceptable accuracy, the gain in time and amplitude of the EC tracings are amplified. In addition, all leads are viewed simultaneously to be able to pick out the earliest Q- or R-wave and the lead with the longest end of T-wave. Notably, the earliest Q-wave and last end of T-wave are shown on different ECG tracings from different leads. In the column grid report format of Figure 9A, the QT interval can be measured on lead II or lead V5 but measuring on only one lead does not allow measurement from earliest Q-wave start to the last T-wave end. In this case, there is only one beat to measure shown in lead V5. In this example, there is no artifact corrupting the one visualized beat. However, it is also possible that there are no visible beats if the heart rate is low enough (long RR interval or cycle time). [0061] The ECG report of Figure 9B shows an example of a custom layout optimized for measuring the QT interval. The gain in time and amplitude is amplified, e.g., above a predetermined threshold. The morphology inside the P, QRS and T-waves can overlap because the QT interval measurement is only concerned with the start and end of the ECG tracing, not the middle. In addition, the noise is reduced by averaging beats together without using filters which can distort the tracings. The earliest QRS start 910 and the latest T-wave end 920 are on different leads. The difference between earliest QRS start 910 and last QRS start 930 is visually different. The difference between earlier T-wave end 940 and later T-wave end 920 is also easily visualized.

[0062] Figures 10A, 10B and IOC show examples of the ECG report for analyzing

T-wave abnormality according to some embodiments. For example, one embodiment measures the degree of T-wave abnormality with the angular difference between QRS and T-wave vector loops in a vector cardiogram. Figure 10A shows a pediatric ECG report where the ECG interpretation suggests right ventricular hypertrophy (RVH).

[0063] Figures 10B and IOC show vectorcardiogram (VCG) type ECG reports to classify hypertrophy and conduction detects. Those ECG reports show two dimensional QRS VCG loop 1010 and/or 1020, as well the magnified T-wave VCG loop 1030 and/or 1040. The angular difference in the QRS and T-wave loops is quite different in this example. Both loops head away from the origin in different directions. The vector loop display can also show the width of the T-wave loop which also shows the degree of repolarization abnormality in adults. A narrow T-wave loop is normal while a very round T-wave loop is more abnormal.

[0064] Figures 11A, 11B and 11C show different examples of QRS and T-wave loops for a normal adult ECG. Figure 11 A shows the ECG report having 3x4 column grid layout. Figures 11B and 11C show the frontal plane and horizontal plane vector loops. In this example, the T-wave loops 1120 and 1140 are narrow and the angles between the QRS loops 1110 and 1130 and corresponding T-wave loops are also narrow.

[0065] Figures 12A and 12B show examples of ECG reports to diagnose acute coronary syndrome (ACS). ACS refers to a group of conditions due to decreased blood flow in the coronary arteries such that part of the heart muscle is unable to function properly or dies. The most common symptom is chest pain, often radiating to the left arm or angle of the jaw, pressure-like in character, and associated with nausea and sweating.

[0066] For evaluating morphology in cases for suspected ACS, rhythm is less important than the morphology of QRS, ST-segment and T-waves. Figure 12A shows the 3x4 column grid layout of the ECG report showing an evolving infarct. The infarct is known to be evolving by comparison to the patient's previous ECG. In the layout of the ECG report of Figure 12B, the morphology is emphasized by magnification in time and amplitude and by averaging to reduce noise. In addition, this layout can overlay corresponding ECG tracing from the current and previous ECG data to show the evolution of the infarct. In this example the ST elevation from the previous ECG data has resolved and the current ECG data show more evidence of recent MI,e.g. the Q-waves in leads III and aVF have increased while the R-wave in V4 is reduced.

[0067] The above-described embodiments can be implemented in any of numerous ways. For example, the embodiments may be implemented using hardware, software or a combination thereof. When implemented in software, the software code can be executed on any suitable processor or collection of processors, whether provided in a single computer or distributed among multiple computers. Such processors may be implemented as integrated circuits, with one or more processors in an integrated circuit component. Though, a processor may be implemented using circuitry in any suitable format.

[0068] Also, the embodiments may be embodied as a method, of which an example has been provided. The acts performed as part of the method may be ordered in any suitable way. Accordingly, embodiments may be constructed in which acts are performed in an order different than illustrated, which may include performing some acts simultaneously, even though shown as sequential acts in illustrative embodiments.

[0069] Use of ordinal terms such as "first," "second," in the claims to modify a claim element does not by itself connote any priority, precedence, or order of one claim element over another or the temporal order in which acts of a method are performed, but are used merely as labels to distinguish one claim element having a certain name from another element having a same name (but for use of the ordinal term) to distinguish the claim elements.

[0070] Although various embodiments have been described by way of examples, it is to be understood that various other adaptations and modifications can be made within the spirit and scope of the present disclosure. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.

Claims

CLAIMS What is claimed is:

1. A method for digital visualization of electrocardiography (ECG) data, comprising: receiving ECG data of electrical activity of a heart of a patient collected over a period of time, wherein the ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes;

calculating one or more ECG statistics of the ECG data;

determining an ECG interpretation using the ECG statistics;

selecting, using a mapping between a set of ECG interpretations and a set of report formats, a report format corresponding to the ECG interpretation;

converting the ECG data according to the report format to produce an ECG report; and

rendering the ECG report to an output interface, wherein steps of the method are performed by a processor.

2. The method of claim 1, wherein the ECG statistics include one or combination of a baseline content, an average amplitude, a peak amplitude, an average frequency, a R- wave count, and an ECG rate of the ECG data.

3. The method of claim 1, further comprising:

determining the ECG interpretation using one or combination of a probabilistic Bayesian analysis, cluster analysis, artificial neural networks, and regression analysis.

4. The method of claim 1, further comprising:

determining the ECG interpretation using a regression function providing a relationship between values of the ECG statistics and the set of ECG interpretations.

5. The method of claim 1, further comprising:

filtering the ECG data to reduce noise of the ECG data before calculating the ECG statistics.

6. The method of claim 1, wherein the output interface includes one or combination of a display interface of a display device, an imaging interface of an imaging device, and a printer interface of a printing device.

7. The method of claim 1, wherein the report format includes a type of the report and a layout of the report, wherein the converting comprises: selecting a subset of ECG tracings from the set of ECG tracings according to the layout of the report;

selecting a duration of the subset of ECG tracings according to the layout of the report;

converting the subset of ECG tracings into a set of digital images according to the type of the report; and

mutually arranging the subset of the images according to the type of the report to form the ECG image.

8. The method of claim 7, wherein the converting amplifies values of at least one ECG tracing.

9. The method of claim 1, further comprising:

determining a custom report layout tailored for a particular ECG interpretation; and updating the mapping with a correspondence between the particular ECG interpretation and the custom report layout.

10. The method of claim 1, further comprising:

receiving ECG signals from a medical device including the set of electrodes configured to detect the ECG signals; and converting the ECG signals into the set of ECG tracings forming the ECG data.

11. A system for digital visualization of electrocardiography (ECG) data, comprising: a memory to store a mapping between a set of ECG interpretations and a set of report formats;

an output interface to render an ECG report;

a network interface to receive ECG data of electrical activity of a heart of a patient collected over a period of time, wherein the ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes; and

a processor configured to

calculate one or more ECG statistics of the ECG data;

determine an ECG interpretation using the ECG statistics;

select, using the mapping, a report format corresponding to the ECG interpretation;

convert the ECG data according to the report format to produce the ECG report; and

render the ECG report to the output interface.

12. The system of claim 11, further comprising: a medical device including a set of electrodes configured to detect ECG signals of the heart of the patient; and

an analog-to-digital converter to covert the ECG signals into the ECG data.

13. The system of claim 11, wherein the report format includes a type of the report and a layout of the report, and wherein the processor is configured for

selecting a subset of ECG channels from the set of ECG channels according to the layout of the report;

selecting a duration of the subset of ECG channels according to the layout of the report;

converting the subset of ECG channels into a set of images according to the type of the report; and

14. The system of claim 11, further comprising an update interface for:

accepting a custom report layout tailored for a particular ECG interpretation; and updating the mapping with a correspondence between the particular ECG interpretation and the custom report layout.

15. A non-transitory computer readable storage medium embodied thereon a program executable by a processor for performing a method, the method comprising:

receiving ECG data of electrical activity of a heart of a patient collected over a period of time, wherein the ECG data include digital representation of a set of ECG tracings recorded by a corresponding set of electrodes;

calculating one or more ECG statistics of the ECG data;

determining an ECG interpretation using the ECG statistics;

rendering the ECG report to an output interface.