ECG training and skill improvement

Abstract

A diagnostic electrocardiogram system employs an electrode lead system (40) to generate one or more electrode signals indicative of electrical activity of a subject's heart (10). The diagnostic electrocardiogram system also employs a diagnostic electrocardiograph (50) coupled to the electrode lead system (40) to communicate (e.g., list, display, and / or print) a subject electrocardiogram (20) and one or more diagnostic electrocardiograms (30) designated as morphologically matching the subject electrocardiogram (20), involving determining a probability that the diagnostic electrocardiogram(s) (30) represent an accurate diagnostic assessment of the subject electrocardiogram (20) based on a similarity between a morphology of the subject electrocardiogram (20) and a morphology of the at least one diagnostic electrocardiogram (30). The subject electrocardiogram (20) provides one or more interpreted information of ECG features derived from the electrical activity of the subject's heart (10) indicated by the electrode signal(s). The diagnostic electrocardiogram(s) provide one or more diagnosed information of ECG features derived from recorded electrical activity of a diagnostic heart (11).

Description

Technical Field

[0001] This disclosure generally relates to electrocardiogram (ECG) training and skill enhancement, and more specifically to systems, apparatus, and methods for ECG training and skill enhancement by morphological matching of ECGs through communication (e.g., display, printing, linking, etc.) from a training ECG set. Background Technology

[0002] The skill of reading n-lead ECGs (e.g., 12-lead ECGs) typically begins with textbook examples and explanations of ECG phenomena. Further skill development usually involves supervised reading of ECGs. ECG skills are typically improved through practice and feedback from experts. At some points, there is no prepared feedback and the ECG operator is alone. Examples of ECG operators include, but are not limited to: physicians, nurse practitioners, physician assistants, nurses, paramedics, trained nursing assistants, and emergency medical technicians.

[0003] More specifically, interpreting or “reading” ECGs is often a skill that requires considerable time and practice to truly master. There exists a general understanding of the technical aspects of ECGs, and most importantly, of how they reflect various cardiac conditions in the ECG signal. ECG training typically begins with textbook explanations of where the signal comes from, how it is recorded, and how signals from the four chambers appear in the ECG signal. Textbook instruction usually includes exemplary ECGs covering arrhythmias and signal morphology in key areas, such as conduction system problems, infarction, and ischemic conditions. Some in-service training typically involves training usually validated by the nursing or medical department.

[0004] The problem is that this training is often inconsistent. Furthermore, ECG operators often do not receive feedback on the quality or accuracy of their ECG interpretations. Additionally, patients frequently have long lists of comorbidities that cause confusion with other effects present on the ECG. Textbook ECG examples almost never include these confounding effects, as they are confusing for novice ECG operators.

[0005] Electrocardiogram (ECG) operators will benefit from being able to find a set of example ECGs that are similar to types they don't frequently see. The problem is that example ECGs are usually organized by ECG interpretation. Therefore, it is necessary to understand existing interpretations to find similar examples.

[0006] ECG is currently the most common cardiac examination performed on patients with suspected cardiac conditions in many settings, including primary care, in the field, or in an ambulance. Although it is accepted as core medical practice, it is believed that only a low percentage of ECG interpreters receive formal training and evaluation in ECG interpretation. In recent years, many ECG interpreters have relied on computer algorithms to interpret their ECGs. However, such algorithms are not perfect, as they often lack access to the clinical context and other necessary information to reliably make accurate diagnoses. This is why it is often mandatory for all computer-interpreted ECGs in clinical settings to be verified and properly corrected by experienced ECG interpreters. More specifically, although many physicians learn the cognitive skills required for proper ECG interpretation, such as during fellowship or residency training, fellowship or residency completion does not guarantee competence. This disclosure can help ECG interpreters continue their on-the-job training and assist with those ECGs that are difficult to interpret. As those skilled in the art will recognize in light of the teachings herein, this disclosure can also have many other benefits. Summary of the Invention

[0007] This disclosure assists electrocardiograph operators (e.g., physicians, nurse practitioners, physician assistants, nurses, paramedics, medical assistants, trained nursing assistants, and emergency medical technicians) in continuously improving their ECG reading skills by providing (and / or offering, displaying, printing, or otherwise communicating) a set of similar ECGs for interpreting or "reading" or otherwise making them available for tracking, storing, processing, etc., in a specific setting. Generally, a relatively large number of ECGs is preferred in the training set.

[0008] According to exemplary embodiments of this disclosure, in the primary application of electronic ECG editing, the invention provides ECG operators with example ECGs similar to the ECG they are currently editing or observing. The invention selects similar ECGs by the characteristics of the signal rather than by correct interpretation. In this way, ECG operators can see many ECGs with similar appearances but potentially different interpretations, as many ECG characteristics have a potentially different set of diagnoses. ECG operators can not only see different diagnostic possibilities, but they can also see the perspectives of different ECG operators on similar ECGs, as the database consists of prior ECGs from their and / or associated (one or more) institutions. Furthermore, the invention can provide the probability that the ECG in question falls into a specific diagnostic category, such as, for example, left bundle branch block (LBBB), right bundle branch block (RBBB), left ventricular hypertrophy, right ventricular hypertrophy, left anterior fascicular block, acute myocardial infarction, early myocardial infarction, and many other conditions. Only the higher probability can be presented to the user.

[0009] One form of the invention disclosed herein is a diagnostic electrocardiogram (ECG) system employing an electrode lead system to generate one or more electrode signals indicative of electrical activity in a subject's heart. The diagnostic ECG system further employs a diagnostic ECGer coupled to the electrode lead system to communicate a subject ECG and one or more diagnostic ECGs (e.g., linking, displaying, and / or printing the morphologically matched subject ECG and(one or more) diagnostic ECGs) determined by the diagnostic ECGer to be morphologically matched to the subject ECG, involving determining the probability that(one or more) diagnostic ECGs (30) accurately represent the diagnostic evaluation of the subject ECG (20) based on the similarity between the morphology of the subject ECG (20) and the morphology of the at least one diagnostic ECG (30). The subject ECG includes one or more interpretations (e.g., algorithmic interpretation of the subject ECG and / or human interpretation by an ECGerman) of ECG features derived from the electrical activity of the subject's heart indicated by the electrode signals(one or more). One or more diagnostic electrocardiograms include one or more diagnoses derived from the recorded electrical activity of the heart (one or more) of the diagnosed heart (e.g., algorithmic diagnosis of one or more diagnostic electrocardiograms and / or electrocardiograph personnel diagnosis).

[0010] The designation of the diagnostic electrocardiograph can be achieved by using a clustering tree that compares the diagnostic electrocardiograph navigation with the training set of diagnostic electrocardiograms. The dimensional space of this clustering tree is derived from linear regression modeling of the ECG features of the training set of diagnostic electrocardiograms.

[0011] A second aspect of the invention disclosed herein is the aforementioned electrocardiograph, which employs a target ECG controller to control the generation of a target ECG. The electrocardiograph also employs a diagnostic ECG controller to control the determination of one or more diagnostic ECGs that match the morphology of the target ECG.

[0012] A third aspect of the invention disclosed herein is a diagnostic electrocardiograph method, which involves the diagnostic electrocardiograph designating at least one diagnostic electrocardiogram (ECG) as a morphological match to a target ECG. This includes the diagnostic ECG determining, based on the similarity between the morphology of the target ECG and the morphology of one or more diagnostic ECGs, the probability that one or more diagnostic ECGs represent an accurate diagnostic assessment of the target ECG. The target ECG provides information on one or more interpretations of ECG features derived from the electrical activity of the target heart indicated by one or more electrode signals generated by a lead system, and the one or more diagnostic ECGs provide information on one or more diagnoses of ECG features derived from the recorded electrical activity of the one or more diagnosed hearts. The diagnostic ECG method further involves the diagnostic ECG communicating the designation of at least one diagnostic ECG as a morphological match to the target ECG (e.g., linking, displaying, and / or printing the target ECG and one or more morphologically matching diagnostic ECGs).

[0013] For the purposes of this disclosure, the term "electrocardiograph" broadly includes all devices known prior to and after this disclosure for recording the electrical activity of the heart over a period of time, and the term "ECG device" broadly includes all individual electrocardiographs and devices / systems that include an electrocardiograph, including, but not limited to:

[0014] (1) Diagnostic ECG equipment (e.g., PageWriter TC ECG machine, Efficia series ECG machine);

[0015] (2) Fitness ECG equipment (e.g., ST80i stress testing system);

[0016] (3) Mobile ECG equipment (Holter monitor);

[0017] (4) Bedside ECG monitoring devices (e.g., IntelliVue monitor, SureSigns monitor, and Goldway monitor);

[0018] (5) Hemodynamic monitoring (e.g., per Flex cardiac physiology monitoring system);

[0019] (6) Telemetry ECG equipment (e.g., IntelliVue MX40 monitor);

[0020] (7) Automated external defibrillators and advanced life support products (e.g., HeartStart MRx and HeartStart XL defibrillators, and Efficia DFM100 defibrillator / monitor);

[0021] (8) An ECG management system (e.g., IntelliSpace ECG management system); and

[0022] (9) Central monitoring system (e.g., PIIC iX and IntelliVue IL central monitoring system).

[0023] Similarly, for the purposes of this disclosure,

[0024] (1) The term “diagnostic electrocardiograph” broadly includes all electrocardiographs having a structural configuration that incorporates the inventive principles of the present disclosure as exemplary as described herein, and the term “diagnostic electrocardiograph method” broadly includes all methods for training and / or operating a diagnostic electrocardiograph that incorporates the inventive principles of the present disclosure as exemplary as described herein.

[0025] (2) Terms in the art, including but not limited to “electrocardiogram operator,” “electrode,” “electrocardiogram,” “ECG feature,” “interpretation,” “diagnosis,” “linear regression,” and “clustering tree,” shall be interpreted as being understood in the field of this disclosure and as being described exemplary herein.

[0026] (3) More specifically with the invention of this disclosure, as understood in the field of this disclosure and as exemplarily described herein, the term “electrocardiogram” broadly includes all types of electrocardiograms used to record the electrical activity of the heart, including but not limited to 12-lead electrocardiograms and 3-lead vectorcardiograms;

[0027] (4) Any descriptive labels used for the term “electrocardiogram” in this document (such as “object electrocardiogram” or “diagnostic electrocardiogram”) serve to distinguish between the electrocardiograms described and claimed herein without specifying or implying any additional limitations on the term “electrocardiogram”.

[0028] (5) More specifically with the invention of this disclosure, as understood in the field of this disclosure and as exemplarily described herein, the term “interpretation” broadly encompasses one or more proposed interpretations of the normality or abnormality of the morphology of an electrocardiogram, as would be understood by those skilled in the art. Examples of electrocardiogram interpretation include, but are not limited to, algorithmic interpretation of electrocardiograms generated by an electrocardiograph and electrocardiograph-person interpretation of electrocardiograms annotated by an electrocardiograph operator;

[0029] (6) More specifically, in the invention of this disclosure, as understood in the field of this disclosure and as exemplarily described herein, the term “diagnosis” broadly encompasses one or more formalized statements of the normality or abnormality of the morphology of an electrocardiogram, as would be understood by those skilled in the art. Examples of electrocardiogram diagnosis include, but are not limited to, the algorithmic interpretation of electrocardiograms generated by an electrocardiograph and the formulation, confirmation, consent, and acceptance of electrocardiograph interpretations of electrocardiograms annotated by an electrocardiograph operator;

[0030] (7) More specifically with the invention of this disclosure, as understood in the field of this disclosure and as exemplarily described herein, the term “cheap ECG features” broadly includes global features and lead-specific features of an electrocardiogram, including but not limited to QRS axis, QRS duration, QT interval, Q / R / S wave amplitude, ST segment amplitude, T wave amplitude, and vector loop.

[0031] (8) More specifically with the invention of this disclosure, as understood in the field of this disclosure and as exemplarily described herein, the term “expensive ECG feature” broadly includes ECG features derived from comparable processing of multiple ECGs, including but not limited to template matching, cross-correlation and RMS differences between ECGs.

[0032] (9) As understood in the field of this disclosure and as exemplarily described herein, the term “feature vector” broadly includes an m-dimensional vector (m≥1) or a vector ring of (one or more) ECG features;

[0033] (10) As understood in the field of this disclosure and as exemplarily described herein, the term “morphological matching” broadly includes the similarity of one or more ECG features between corresponding electrode signals of a pair of electrocardiograms, wherein the one or more ECG features are characteristics of the shape of the electrocardiogram.

[0034] (11) As understood in the field of this disclosure and as exemplarily described herein, the term “diagnostic category” broadly encompasses categories that represent specific diagnostic assessments of an electrocardiogram. Examples of diagnostic categories include, but are not limited to: (a) ventricular conduction defects, including interpretations of left anterior fascicular block, left bundle branch block (LBBB), and right bundle branch block (RBBB); (b) hypertrophy, including interpretations of left ventricular hypertrophy and right ventricular hypertrophy; and (c) ischemia and infarction, including interpretations of acute myocardial infarction, early myocardial infarction, and subendocardial ischemia.

[0035] (12) As exemplarily described herein, the term “accurate diagnostic probability” broadly encompasses the probability of an accurate diagnostic assessment of an electrocardiogram for a specific diagnostic category.

[0036] (13) As understood in the field of this disclosure and as exemplarily described herein, the term “controller” broadly encompasses all structural configurations of a dedicated motherboard or application-specific integrated circuit housed within or linked to an electrocardiograph to control the application of the various inventive principles of this disclosure as subsequently described herein. The structural configuration of a controller may include, but is not limited to: one or more processors, one or more computer-usable / computer-readable storage media, one or more operating systems, one or more application modules, one or more peripheral controllers, one or more slots, and one or more ports. Any descriptive labels used to describe a controller herein (e.g., “object ECG” controller and “diagnostic ECG” controller) are intended to identify the specific controller described and claimed herein without specifying or implying any additional limitations on the term “controller”;

[0037] (14) The term “application module” broadly includes components of a controller, including circuitry and / or executable programs (e.g., executable software and / or firmware stored on one or more non-transient computer-readable media) for performing a particular application. Any descriptive labels of application modules herein (e.g., “ECG feature extractor” module and “clustering tree generator” module) are used to identify specific application modules as described and claimed herein without specifying or implying any additional limitations on the term “application module”;

[0038] (15) The term “communication” broadly includes all communication schemes known prior to, concurrent with, and after this disclosure that are used by an electrocardiograph to communicate an electrocardiogram (ECG) to a user of the ECG. Examples of such communication schemes include, but are not limited to: providing a link to the ECG, displaying the ECG, and printing the ECG;

[0039] (16) As understood in the field of this disclosure and as exemplarily described herein, the terms “signal” and “data” broadly encompass all forms of detectable physical quantities or impulses (e.g., voltage, current, or magnetic field strength) used to transmit information in support of the various inventive principles of this disclosure as subsequently described herein.

[0040] (17) Any descriptive label used herein for the term “signal” serves to distinguish between signals as described herein and those claimed, without specifying or implying any additional limitation on the term “signal”; and

[0041] (18) Any descriptive label used for the term “data” in this document facilitates the distinction between data as described herein and data for which protection is claimed, without specifying or implying any additional limitation on the term “data”.

[0042] The foregoing and other forms of this disclosure, as well as its various features and advantages, will become clearer from the following detailed description of various embodiments, taken in conjunction with the accompanying drawings. The detailed description and drawings are merely illustrative and not limiting of this disclosure, the scope of which is defined by the claims and their equivalents. Attached Figure Description

[0043] Figure 1 The illustration shows exemplary embodiments of object electrocardiograms and diagnostic electrocardiograms based on the inventive principles of this disclosure.

[0044] Figure 2 The illustration shows an object ECG and a pair of exemplary embodiments for diagnosing ECGs according to the inventive principles of this disclosure.

[0045] Figure 3 The illustration shows an object ECG and a pair of exemplary embodiments for diagnosing ECGs according to the inventive principles of this disclosure.

[0046] Figure 4 An exemplary embodiment of a diagnostic electrocardiograph based on the inventive principles of this disclosure is illustrated.

[0047] Figure 5 An exemplary embodiment of a diagnostic ECG controller based on the inventive principles of this disclosure is illustrated.

[0048] Figure 6 The illustration shows a flowchart of an exemplary embodiment of a diagnostic electrocardiograph training method according to the inventive principles of this disclosure.

[0049] Figure 7 An exemplary embodiment of generating a vector ring version of the ECG feature vector according to the inventive principles of this disclosure is illustrated.

[0050] Figure 8A and Figure 8B An exemplary embodiment of the construction of a clustering tree according to the inventive principles of this disclosure is illustrated.

[0051] Figure 9 The illustration shows a flowchart of an exemplary embodiment of a diagnostic electrocardiograph operating method according to the inventive principles of this disclosure. Detailed Implementation

[0052] For ease of understanding of this disclosure, compared to electrocardiograms of objects known in the art of this disclosure, Figure 1The following description teaches the inventive principles of the diagnostic electrocardiogram of this disclosure. More specifically, this disclosure is based on one or more diagnostic electrocardiograms specified by an electrocardiograph that match the morphology of a target electrocardiogram, wherein the target electrocardiogram is generated based on current ECG monitoring and / or testing of the target heart by the electrocardiograph, and wherein the one or more diagnostic electrocardiograms are generated from previous diagnostic ECG monitoring and / or testing of one or more non-target hearts (i.e., one or more diagnosed hearts). Figure 1 From the description, those skilled in the art will recognize how the inventive principles of this disclosure can be applied to make and use the diagnostic electrocardiograph of this disclosure in many and various embodiments.

[0053] refer to Figure 1 The object ECG 20 is an example of an object ECG that can be communicated (e.g., displayed, printed, linked, etc.) by an electrocardiograph during ECG monitoring and / or testing of the object heart 10 via any type of lead system known in the art of this disclosure (e.g., 12-lead system, 3-lead system, etc.). The object ECG 20, as communicated by the electrocardiograph, includes (one or more) graphic images 21, such as, for example, a 12-lead ECG 22 known in the art of this disclosure, an ECG waveform 23 generated as known in the art of this disclosure, and an ECG vector map (not shown) known in the art of this disclosure. The exemplary object ECG 21 also includes a textual interpretation 24 of the normality or abnormality of the ECG morphology of (one or more) graphic images 21, said graphic images 21 being generated by an interpretation algorithm performed by a diagnostic electrocardiograph known in the art of this disclosure and / or by annotated interpretation by an electrocardiograph operator via a graphical user interface of an associated electrocardiograph known in the art of this disclosure. More specifically, interpretation 24 involves one or more interpretations of the normality or abnormality of the ECG morphology of (one or more) graphic images 21, as will be understood by those skilled in the art.

[0054] Still referencing Figure 1 The diagnostic ECG 30 is an example of X number (X≥1) of diagnostic ECGs that can be communicated (e.g., displayed or printed) by an electrocardiograph during the monitoring and / or testing of the aforementioned target heart 10. Each diagnostic ECG 30 is generated from previous ECG monitoring and / or testing of a non-target heart 11 (i.e., the diagnosed heart) via any type of lead system known in the art of this disclosure (e.g., a 12-lead system, a 3-lead system, etc.).

[0055] Each diagnostic ECG 30, as communicated by an electrocardiograph, includes (one or more) graphic images 31, such as, for example, a 12-lead ECG 32 known in the art of this disclosure, an ECG waveform 33 generated as known in the art of this disclosure, and an ECG vector map (not shown) known in the art of this disclosure. The exemplary diagnostic ECG 31 also includes a textual diagnosis 34 by an ECG operator regarding the normality or abnormality of the ECG morphology of (one or more) graphic images 31. Each ECG diagnosis 34 relates to one or more formal interpretations of the normality and / or abnormality of the ECG morphology of the corresponding graphic images 31, as would be understood by someone skilled in the art (e.g., the formulation, confirmation, consent, acceptance, etc., of the interpretation of the diagnostic ECG).

[0056] Still referencing Figure 1 By simultaneously communicating the target electrocardiogram (ECG) and one or more diagnostic ECGs designated as morphological matches to the target ECG, this disclosure improves the capabilities of an electrocardiograph to facilitate accurate diagnosis of the target ECG by an electrocardiograph operator.

[0057] For example, Figure 2 The illustration shows a demonstrative 12-lead ECG 22a with QRS morphology in leads V1 through V4, which can be interpreted algorithmically and / or via annotation as left bundle branch block, left ventricular hypertrophy, or early myocardial infarction. These possible interpretations of the 12-lead ECG 22a make it difficult for ECG operators, especially inexperienced ones, to make a diagnosis based on its morphology.

[0058] Figure 2 The diagram also illustrates a pair of 12-lead ECGs 32(1) and 32(2) from a sample database of approximately 10,000 diagnostic ECGs, designated as morphological matches by the electrocardiograph because the ECG morphology of the QRS in leads V1 to V4 of 12-lead ECGs 32(1) and 32(2) is substantially identical to that of the ECG morphology of the QRS in leads V1 to V4 of 12-lead ECG 22a. Based on these morphological matches, diagnoses of, for example, left bundle branch block in 12-lead ECG 32(1) and 12-lead ECG 32(2) of 12-lead ECG 22a improve the ability of electrocardiograph operators to give accurate diagnoses of left bundle branch block based on the ECG morphology of the QRS in leads V1 to V4 of 12-lead ECG 22a.

[0059] Through another example, Figure 3The illustration shows a demonstrative 12-lead ECG 22b of a target ECG morphology, which can be interpreted algorithmically and / or via annotation as right bundle branch block with or without ischemia (i.e., ST segment inhibition and inverted T waves). These possible interpretations of the 12-lead ECG 22b make it difficult for ECG operators, especially inexperienced ones, to make a diagnosis of the morphology of the 12-lead ECG 22b.

[0060] Figure 3 The diagram also illustrates a pair of 12-lead ECGs 32(3) and 32(4) from a sample database of approximately 10,000 diagnostic ECGs. These were designated as morphological matches by the electrocardiograph because the abnormal morphology of the QRS in leads V1–V4 of 12-lead ECGs 32(3) and 32(4) was substantially identical to the ECG morphology of the QRS in leads V1–V4 of 12-lead ECG 22b. Based on these morphological matches, the diagnosis of, for example, ischemic right bundle branch block using 12-lead ECGs 32(3) and 32(4), improves the ability of electrocardiograph operators to give an accurate diagnosis of ischemic right bundle branch block based on the ECG morphology of the QRS in leads V1–V4 of 12-lead ECG 22b.

[0061] As those skilled in the art would know Figure 1 The teachings should be recognized that, regardless of the interpretation of the subject's electrocardiogram, this disclosure provides confidence to the electrocardiogram operator in giving a diagnosis of the subject's electrocardiogram when the interpretation of the subject's electrocardiogram characteristically matches the diagnosis of one or more diagnostic electrocardiograms.

[0062] To further facilitate understanding of this disclosure, Figure 4 The following description teaches the inventive principles of the diagnostic electrocardiograph of this disclosure. Based on this description, those skilled in the art will recognize how the inventive principles of this disclosure can be applied to make and use many and various embodiments of the diagnostic electrocardiograph of this disclosure.

[0063] refer to Figure 4 The diagnostic electrocardiograph 50 of this disclosure employs a control network 60, a display 70, one or more user input devices 80 (e.g., one or more buttons, one or more dials, touchpads, etc.) and a printer 90. The diagnostic electrocardiograph 50 may also employ one or more additional devices as known in the art of this disclosure (e.g., speakers and LED status indicators).

[0064] The diagnostic electrocardiograph 50 is linked to and / or includes any necessary hardware / software interface to the cable connector 40 to receive one or more electrode signals from the electrode lead system connected to the subject 12 in order to monitor and / or test the subject's heart 10 (e.g., a standard 12-lead connection, such as the Mason-Likar lead system as shown, or a reduced lead system, such as the EASI lead system).

[0065] Control network 60 includes an object ECG controller 61, a diagnostic ECG controller 62, an ECG display controller 63, and an ECG printer controller 64, which are linked to or housed within the diagnostic ECG machine 50 as shown. In practice, controllers 61-64 may be integrated to the extent of design and / or separated as shown. Similarly, in practice, control network 60 may include one or more additional controllers known in the art of this disclosure (e.g., canopy controllers, automated defibrillator controllers, etc.).

[0066] The object ECG controller 61 is configured, as known in the art of this disclosure, to control the generation of an object ECG based on one or more electrode signals (e.g., object ECG controllers commercially used in Holter monitors, IntelliVuem monitors, HeartStart MRx defibrillators, and HeartStart XL defibrillators). In practice, the generation of the object ECG by the object ECG controller 61 includes one or more object ECG graphic images (e.g., Figure 1 The generation of (one or more) object graphic ECG images 21, and may also include algorithmic generation of one or more interpretations of (one or more) object ECG graphic images and / or ECG personnel annotation (e.g., Figure 1 Interpretation of (one or more) objects (24).

[0067] The diagnostic ECG controller 62 is configured according to the inventive principles of this disclosure to transmit one or more diagnostic electrocardiograms (e.g., Figure 1 The diagnostic electrocardiogram (30) is specified as a morphological match with the target electrocardiogram, as will be combined in this paper. Figure 5-9 Further exemplary description.

[0068] The ECG display controller 63 is configured, as known in the art of this disclosure, to display an electrocardiogram (ECG) (e.g., an ECG display controller commercially used in Holter monitors, IntelliVuem monitors, HeartStart MRx defibrillators, and HeartStart XL defibrillators) and to display a graphical user interface for accessing the diagnostic ECG according to the inventive principles of this disclosure. In practice, the display of an ECG via the ECG display controller 63 may include:

[0069] 1. User customization of the view of an object ECG graphic image via one or more user input devices 80 and / or a graphical user interface (not shown);

[0070] 2. User annotations for algorithmic interpretation of the object's electrocardiogram via one or more user input devices 80 and / or a graphical user interface (not shown); and / or

[0071] 3. User selection of a diagnostic electrocardiogram (ECG) displaying a grid of large thumbnail images of the diagnostic ECG (“ECG grid”), a label organization of the diagnostic ECG (“ECG label”), or any other icon suitable for the management review of the diagnostic ECG, via one or more user input devices 80 or the diagnostic graphical user interface 25.

[0072] ECG printer controller 64 is configured, as known in the art of this disclosure, to operate printer 90 to print electrocardiograms via one or more user input devices 80 and / or graphical user interfaces (not shown) (e.g., ECG printer controllers commercially used by Holter monitors, IntelliVuem monitors, HeartStart MRx defibrillators, and HeartStart XL defibrillators).

[0073] For a better understanding of this disclosure, Figure 5 The following description teaches the inventive principles of the diagnostic ECG controller of this disclosure. Based on this description, those skilled in the art will recognize how the inventive principles of this disclosure can be applied to make and use many and various embodiments of the diagnostic ECG controller of this disclosure.

[0074] refer to Figure 5 Diagnostic ECG controller 62 ( Figure 4 Example 62a employs an ECG feature extractor 100, an ECG profile builder 110, and a clustering tree builder 120 for the purpose of training the diagnostic ECG controller 62a, as will be discussed herein. Figure 6 -8 further describes this. The diagnostic ECG controller 62 also employs a clustering tree navigator 130, an ECG morphology matcher 140, and a diagnostic category assigner 150 to operate the diagnostic ECG controller 62a for monitoring / testing purposes, as will be further described herein in conjunction with Figures 8 and 9. For both training and monitoring / testing purposes, the diagnostic ECG controller 62a may also employ a database manager 160 and a diagnostic ECG database 170 as shown, or alternatively communicate with the database manager 160 for the purpose of accessing the diagnostic ECG database 170.

[0075] The diagnostic ECG database 170 stores a number of diagnostic ECGs 30 as shown in Figure X. As previously described herein, each diagnostic ECG 30 is generated based on previous diagnostic ECG monitoring and / or testing of a non-target heart (i.e., the heart being diagnosed). Each diagnostic ECG 30 includes one or more graphic images, such as, for example, a 12-lead ECG, ECG waveforms, and / or ECG vectorgraphs. Each diagnostic ECG 30 also includes an ECG diagnosis of the ECG morphology of the one or more graphic images by an electrocardiograph operator, wherein each ECG diagnosis involves one or more formal interpretations of the ECG morphology of one or more corresponding graphic images 31 by an electrocardiograph operator, as will be understood by those skilled in the art.

[0076] Still referencing Figure 5 The ECG feature extractor 100 is configured with hardware, software, firmware, and / or circuitry for processing electrocardiograms, objects, or diagnoses to calculate a cheap ECG feature vector (“IEFV”) 101 from the electrocardiogram, wherein the IEFV 101 comprises m cheap ECG features (m ≥ 1). Examples of cheap ECG features include, but are not limited to, QRS axes, QRS duration, QT interval, Q / R / S wave amplitude, ST segment amplitude, T wave amplitude, and vector loop. In practice, the ECG feature extractor 100 may implement any technique known in the art for cheap ECG features as disclosed herein.

[0077] The ECG feature extractor 100 is also configured with hardware, software, firmware, and / or circuitry for processing ECG, object-diagnosis, and / or diagnosis-diagnosis pairings and / or for processing inexpensive ECG feature vectors 101, object-diagnosis, or diagnosis-diagnosis pairings to compute expensive ECG feature vectors (“EEFV”) 102 between ECG pairs, wherein the EEFV 101 comprises q numbers of expensive ECG features, q ≥ 1. Examples of expensive ECG features include, but are not limited to, template matching, cross-correlation, and RMS difference between ECG pairs. In practice, the ECG feature extractor 100 may implement any technique known in the art of this disclosure for computing expensive ECG features.

[0078] The ECG diagnostic profiler 110 is configured with hardware, software, firmware, and / or circuitry for processing cheap ECG feature vectors 101 for each diagnostic ECG 30 and each paired expensive ECG feature vector 102 for each ECG 30 to construct a diagnostic ECG profile vector (“DEPV”) 111, said diagnostic ECG profile vector (“DEPV”) 111 comprising n numbers of cheap ECG features, m ≥ n ≥ 1 (i.e., diagnostics among expensive ECG features), that best represent the interpretability of expensive ECG features known in the art of this disclosure. In practice, the ECG diagnostic profiler 110 may implement any techniques for determining which cheap ECG features best model expensive ECG features known in the art of this disclosure, including but not limited to linear regression of IEFV 101 and EEFV 102.

[0079] The clustering tree builder 120 is configured with hardware, software, firmware, and / or circuitry for processing the diagnostic ECG contour vector 111 to construct a clustering tree (“CT”) 121 of nodes and leaves built from the outlined inexpensive ECG features. Each node will be associated with one of the outlined inexpensive ECG features and a corresponding threshold. Each leaf will be associated with one or more diagnostic ECGs 30. In practice, the clustering tree builder 120 may implement any technique for constructing the clustering tree 121, including, but not limited to, constructing a decision tree of the partitioned data space derived from the diagnostic ECG contour vector 111 as a clustered (or dense) region or an empty (or sparse) region formed by partitioned clusters or hierarchical clusters.

[0080] Clustering tree navigator 130 is configured with cheap ECG feature vector 101 for processing object ECGs to navigate the nodes of clustering tree 121 until a leaf is reached. Thus, clustering tree navigator 130 generates a nearest neighbor list (“NNL”) 131 of all(one or more) diagnostic ECGs 30 associated with the reached leaf.

[0081] The ECG morphology matcher 140 is configured with hardware, software, firmware, and / or circuitry for processing the nearest neighbor list 131 to designate one or more of the nearest neighbor diagnostic ECGs 30 as morphological matches to the object ECG. The ECG morphology matcher 140 thereby generates a morphology match list (“EMML”) 141 for each designated nearest neighbor diagnostic ECG 30. In practice, the ECG morphology matcher 140 may implement any known techniques for determining any similarity in ECG morphology between the object ECG and each nearest neighbor diagnostic ECG.

[0082] The diagnostic category assigner 150 is configured with hardware, software, firmware, and / or circuitry for processing the morphological matching list 141 to assign each morphologically matched diagnostic electrocardiogram to one of a plurality of diagnostic categories, wherein each diagnostic category represents a specific diagnostic evaluation of the diagnostic electrocardiogram. Examples of diagnostic categories include, but are not limited to: left bundle branch block (LBBB), right bundle branch block (RBBB), left ventricular hypertrophy, right ventricular hypertrophy, left anterior fascicular block, acute myocardial infarction, and early myocardial infarction.

[0083] Diagnostic category assigner 150 generates a list of diagnostic categories (“DCA”) and associated diagnostic electrocardiograms for each diagnostic category to provide a diagnostic evaluation of the subject’s electrocardiogram. In practice, diagnostic category assigner 150 may also determine the probability that each listed diagnostic category represents an accurate diagnostic evaluation of the subject’s electrocardiogram.

[0084] For a better understanding of this disclosure, Figure 6 The following description teaches the inventive principles of the diagnostic electrocardiograph training method of this disclosure. Based on this description, those skilled in the art will recognize how the inventive principles of this disclosure can be applied to create and use many and various embodiments of the diagnostic electrocardiograph training method of this disclosure.

[0085] refer to Figure 6 Flowchart 200 shows the diagnostic process for ECG controller 62a ( Figure 5 The diagnostic electrocardiograph training method of this disclosure is performed during the training phase. In practice, the diagnostic ECG database 170 will typically include thousands or even millions of large morphological variations of the diagnostic ECG 30. Flowchart 200 facilitates the segmentation of inexpensive ECG features of the diagnostic ECG 30 to best represent the possible interpretation of the morphology of the diagnostic ECG 30.

[0086] Still referencing Figure 6 Phase S202 of flowchart 200 includes an ECG feature extractor 100 that processes a set of diagnostic ECGs 30 to compute a number of cheap ECG feature vectors 101 (X≥Y≥1, (e.g., Y-dimensional vectors of cheap ECG features or alternatively vector loops)) and a number of expensive ECG feature vectors 102 (Z≥1).

[0087] More specifically for 12-lead ECGs, in practice, the cheap ECG feature vector 101 will consist of a processed version of the lead signals rather than a set of measurements (e.g., R-wave amplitude and QRS duration). Since the number of points in the representative beats (or the average beats consisting of beats of similar shape, excluding noise and abnormal beats) can be very large for the cheap feature vector (e.g., 500 points per lead for 12 leads), the number of points should be reduced if possible. This implementation can use (2) methods to reduce the number of points in the cheap ECG feature vector while still preserving the unique morphological information. First, the number of points will be reduced by using a 12-lead ECG to Frank lead vectorcardiogram transformation from 12 leads containing a lot of redundant information to three (3) orthogonal leads. This is a 4:1 reduction in points. Second, the number of points will be further reduced by using an approximation given by a multi-level wavelet decomposition. Using an approximation from a fourth-level decomposition, the final number of points in the cheap ECG feature vector is reduced to approximately 100.

[0088] Figure 7 The illustration shows an exemplary transformation from the representative beats 103 of lead 12 of the inexpensive ECG feature vector 101 to the representative beats 102a of lead Frank. The X, Y, and Z signals of lead Frank are used to generate a two-dimensional vector loop based on the X, Y, and Z signal pairs. In practice, multiple pairs of ECG vectorograms can also be used to generate the expensive ECG feature vector 102.

[0089] In practice, for stage S202, the entire database 170 of the diagnostic electrocardiogram 30 or a subset thereof can be processed by the ECG feature extractor 100 based on various factors.

[0090] For example, the computation of expensive ECG feature vector 102 may in practice involve: a sample of comparisons between each diagnostic ECG 30 and each other ECG 30, or a random sample of a subset of diagnostic ECGs 30, or a target group of diagnostic ECGs 30 expected to be within the same diagnostic group.

[0091] Additionally, if the diagnostic ECG database 70 is relative to the diagnostic ECG controller 62a ( Figure 5 The processing power of the ECG is relatively large, so in practice, the ECG 30 can be segmented by age group and / or gender to limit the size of the resulting cluster tree.

[0092] Furthermore, the diagnostic ECG 30 processed by the ECG feature extractor 100 can, in practice, be based solely on selected ECG operators with years of experience or proven excellence in ECG reading accuracy. This eliminates the need for diagnostic ECG 30 from less experienced ECG operators.

[0093] Furthermore, those skilled in the art will recognize that the ECG morphology of a stress test on the heart of a subject differs from the morphology of a resting diagnostic ECG of the same heart. However, this disclosure is equally applicable to relaxation monitoring and stress testing of the same heart. Therefore, in practice, the diagnostic ECG database 70 can be divided into a resting ECG training database leading to a resting ECG clustering tree and a stress test training database leading to a stress test ECG clustering tree.

[0094] Still referencing Figure 6 Phase S204 of flowchart 200 includes processing Y number of cheap ECG feature vectors 101 and Z number of expensive ECG feature vectors 102 to establish an ECG diagnostic profiler 110 for a diagnostic ECG profile vector (“DEPV”) 111, the diagnostic ECG profile vector (“DEPV”) including n number of cheap ECG features that best represent the interpretability of expensive ECG features.

[0095] In one embodiment of stage S204, the ECG diagnostic profiler 110 performs linear regression or another similar method to determine which cheap ECG feature best models the expensive ECG feature. For this embodiment, the dependent variable is the expensive ECG feature, and the independent variable is the difference in the cheap ECG feature. The training set used for this linear regression operation is the set of differences in the ECG features of each diagnostic ECG 30 compared to the ECG features of other diagnostic ECGs 30 in the training set. In the simplest case, linear regression is performed by fitting a line to a scatter plot of points with one dependent variable and multiple independent variables. After fitting the line to the data (i.e., training), the dependent variable is a linear function of the independent variables or features. The model formula is as follows [1]:

[0096] Y=b0+b1*x1+b2*x2+…+bn*xn, [1]

[0097] Where Y is the dependent variable,

[0098] Where x1, x2, ..., xn are independent variables, and

[0099] b0, b1, ..., bn are coefficients determined during the training operation.

[0100] In the extreme case, the set of rows (each row being a test and each column being a feature) is a comparison of each diagnostic ECG 30 with each other diagnostic ECG 30.

[0101] After the linear regression model is computed, the ECG diagnostic profiler 110 will generate a vector of cheap ECG features with low p-values ​​(i.e., one or more cheap ECG features that make a significant contribution to the dependent variable, as those skilled in the art will recognize).

[0102] Still referencing Figure 6 Phase S206 of flowchart 200 includes a clustering tree generator 120 that processes ECG contour vectors 111 to construct clustering tree 121.

[0103] In one embodiment of stage S206, the clustering tree generator 120 implements a nearest neighbor algorithm with a kd-tree representing a k-dimensional tree. K-dimensionality means there are k features used in the clustering operation. This is a binary tree. Each node in the tree has two nodes: a left node and a right node. These nodes are further nodes that are thus divided into left and right subtrees, representing the left and right results. The terminating (leaf) branch of the tree is a k-dimensional data point. The left and right subtrees represent all points below that are separated by a hyperplane. Because there is k-dimensionality, this is generally a hyperplane. As you move down the tree horizontally from the root node at the top, the separation at each level corresponds to a separation based on exactly one of the k features. Typically, the separation occurs near the middle of that feature. All points in the subtree with a value higher than the median of the subtree are on one side of the hyperplane, and all other points are on the other side. Moving down the tree horizontally, the separation rotates through the features, meaning the separation at the root node is based on the first feature, the separation at the next level uses the next feature, and so on.

[0104] Figure 8A A clustering paradigm is shown for stage S206 involving twenty (20) diagnostic ECGs 30, whereby three (3) diagnostic cheap ECG features DIEF have been determined to be the best interpreters of expensive ECG features (e.g., QRS axis, QRS duration, and QT interval). The diagnostic cheap ECG features DIEF of the twenty (20) diagnostic ECGs 30 are clustered within a three-dimensional data space 123, and the resulting clusters are applied by a clustering tree generator 120 to generate a partitioned data space 124 via feature partitioning units FP(1) to FP(3) (e.g., partitioning units based on the median or pattern of each diagnostic cheap ECG feature).

[0105] Figure 8B The construction of a clustering decision tree 121a based on the partitioned data space 124 is shown. The clustering decision tree 121a includes nodes N1 to N7 and leaves L1 to L8. Node N1 is associated with the diagnostic cheap ECG feature DIEF(1) having a median or pattern value r. Nodes N2 to N3 are associated with the diagnostic cheap ECG feature DIEF(2) having a median or pattern value s. Nodes N4 to N7 are associated with the diagnostic cheap ECG feature DIEF(3) having a median or pattern value t.

[0106] Each leaf with twenty (20) diagnostic electrocardiograms 30 ( Figure 7One or more of A) are associated. For a simple example, leaf L1 can be associated with diagnostic ECGs 30(1) through 30(3). Leaf L2 can be associated with diagnostic ECGs 30(4) and 30(5). Leaf L3 can be associated with diagnostic ECGs 30(6) through 30(9). Leaf L4 can be associated with diagnostic ECG 30(10). Leaf L5 can be associated with diagnostic ECGs 30(11) and 30(12). Leaf L6 can be associated with diagnostic ECGs 30(13) through 30(15). Leaf L7 can be associated with diagnostic ECGs 30(16) through 30(18). Leaf L8 can be associated with diagnostic ECGs 30(19) and 30(20).

[0107] Those skilled in the art will recognize that flowchart 200 will typically involve the processing of thousands (if not millions) of diagnostic electrocardiograms 30, and provide Figure 8A and Figure 8B To demonstrate a simple example to facilitate understanding of stage S206.

[0108] For a better understanding of this disclosure, Figure 9 The following description teaches the inventive principles of the diagnostic electrocardiogram (ECG) operation method of this disclosure. Based on this description, those skilled in the art will recognize how the inventive principles of this disclosure can be applied to make and use many and various embodiments of the diagnostic ECG operation method of this disclosure.

[0109] refer to Figure 9 Flowchart 210 shows the diagnostic process for ECG controller 62a ( Figure 5 The diagnostic electrocardiogram evaluation method of this disclosure is performed during the activation phase of the process. Flowchart 210 facilitates the diagnostic evaluation of the subject's electrocardiogram.

[0110] Still referencing Figure 9 Stage S212 of flowchart 210 includes ECG feature extractor 100, which processes the object's electrocardiogram (e.g., in...). Figure 1 The object ECG 20) and diagnostic ECG contour vector 111 shown are used to generate a cheap ECG feature vector 101s corresponding to the diagnostic ECG contour vector 111. For example, in Figure 8A and Figure 8B In the context of this, ECG feature extractor 100 generates cheap ECG feature vectors 101s including diagnostic cheap ECG features DIEF(1) to DIEF(3).

[0111] Phase S214 of flowchart 210 includes a clustering tree navigator 130, which processes cheap ECG feature vectors 101s to navigate the nodes of clustering tree 121 until a leaf is reached, thereby generating a nearest neighbor list (“NNL”) 131 of all(one or more) diagnostic ECGs 30 associated with the reached leaf. For example, in Figure 8A and Figure 8B In this context, the clustering tree navigator 130 can reach leaf L1 and generate a nearest neighbor list 131 including diagnostic electrocardiograms 30(1) to 30(3).

[0112] Phase S216 of flowchart 210 includes an ECG morphology matcher 140 that processes a nearest neighbor list 131 to generate a morphology match list (“EMML”) 141 for each specified nearest neighbor diagnostic ECG 30.

[0113] In one embodiment of stage S216, the ECG morphology matcher 140 calculates expensive ECG features (e.g., template matching, cross-correlation, or RMS error) between the target ECG 20 and each nearest-neighbor diagnostic ECG 30, and determines the cross-correlation between the mean beats of the target ECG 20 and the mean beats of the nearest-neighbor diagnostic ECGs, resulting in a vector of the number of cross-correlations. The ECG morphology matcher 140 selects a subset of the nearest-neighbor diagnostic ECGs 30 by classifying the cross-correlation vectors from highest to lowest and selecting the subset with one or more of the highest cross-correlations (i.e., most similar to the target ECG 20).

[0114] For example, in Figure 8A and Figure 8B In this context, ECG morphology matcher 140 can designate diagnostic electrocardiograms 30(1) and 30(2) as morphology matches.

[0115] Phase S218 of flowchart 210 includes a diagnostic category assigner that processes the morphological matching list 141 to assign each matched diagnostic ECG to a diagnostic category (where each diagnostic category represents a specific diagnostic assessment of the diagnostic ECG) and determines the probability that each listed diagnostic category represents an accurate diagnostic assessment of the object ECG.

[0116] In one embodiment of stage S218, the probability of a diagnostic category is calculated as the frequency of annotations for that diagnostic category within a subset of nearest-neighbor morphological matches. Specifically, each nearest-neighbor diagnosis for a morphological match is mapped to a broader diagnostic category. The number of times each diagnostic category is annotated is divided by the number of diagnostic ECGs in the subset of nearest-neighbor morphological matches. This ratio is an estimate of the probability.

[0117] For example, in Figure 8Aand Figure 8B In this context, diagnostic ECGs 30(1) and 30(2) can be mapped to left bundle branch block, and diagnostic ECG 30(2) can also be mapped to left ventricular hypertrophy. Therefore, the probability that the subject ECG may show left bundle branch block will be 66%, and the probability that the subject ECG may show left ventricular hypertrophy will be 33%.

[0118] After flowchart 210 ends, the set of nearest neighbor morphological matches is presented to the ECG operator in some other graphical way, such as a grid of large thumbnail images, labeled, or allowing rapid switching from one diagnostic ECG to the next so as to quickly review all diagnostic ECGs for a subset of nearest neighbor morphological matches.

[0119] refer to Figure 1-9 Those skilled in the art will recognize that many benefits of the invention disclosed herein include, but are not limited to, improvements in the diagnostic evaluation of electrocardiograms of subjects.

[0120] The present disclosure has been described herein with reference to preferred embodiments. Modifications or alternatives may arise in the reader’s reading and understanding of the specific embodiments described above. This invention is intended to be construed as including all such modifications and alternatives, provided they fall within the scope of the claims and their equivalents.

[0121] Furthermore, as those skilled in the art will recognize from the teachings provided herein, the features, elements, components, etc., described in this disclosure / specification and / or depicted in the drawings can be implemented as various combinations of electronic components / circuits, hardware, executable software, and executable firmware, and provide functionality that can be combined in a single element or multiple elements. For example, the functionality of the various features, elements, components, etc., shown / illustrated / depicted in the drawings can be provided by using dedicated hardware and hardware capable of being associated with software running suitable software. When provided by a processor, the functionality can be provided by a single dedicated processor, a single shared processor, or multiple individual processors (some of which may be shared and / or multiplexed). Furthermore, the explicit use of the terms “processor” or “controller” should not be construed as referring exclusively to hardware capable of running software, and may implicitly include, but is not limited to: digital signal processor (“DSP”) hardware, memory (e.g., read-only memory (“ROM”), random access memory (“RAM”), non-volatile memory, etc.) for storing software, and any means and / or machine (including hardware, software, firmware, circuitry, combinations thereof, etc.) that are substantially capable of (and / or configurable to) perform and / or control processes.

[0122] Furthermore, all descriptions of the principles, aspects, exemplary embodiments, and specific examples thereof detailed herein are intended to cover structural and functional equivalences. Additionally, it is intended that such equivalences include both currently known equivalences and future developmental equivalences (e.g., any element developed capable of performing the same or substantially similar functions, regardless of its structure). Therefore, for example, in view of the teachings provided herein, those skilled in the art will recognize that any block diagram provided herein can represent a conceptual view of exemplary system components and / or circuits implementing the principles of the invention. Similarly, in view of the teachings provided herein, those skilled in the art will recognize that any flowchart, operation diagram, etc., can represent various processes that can be substantially represented in a computer-readable storage medium and thus operated by a computer, processor, or other device with processing capabilities, whether or not such a computer or processor is explicitly shown.

[0123] Preferred and exemplary embodiments of diagnostic electrocardiographs and their operation methods have been described (these embodiments are intended to be illustrative and not limiting). It is worth noting that those skilled in the art will find the teachings provided herein (including...) helpful. Figure 1-9 Modifications and variations can be made based on the inspiration provided. Therefore, it should be understood that modifications can be made to the disclosed preferred and exemplary embodiments, which are within the scope of the embodiments disclosed herein.

[0124] Furthermore, it should be contemplated that incorporating and / or implementing the aforementioned device, or corresponding and / or related systems such as those that can be used / implemented in the device according to this disclosure, is also contemplated and considered to be within the scope of protection of this invention. Additionally, corresponding and / or related methods for manufacturing and / or using the device and / or system according to this disclosure are also contemplated and considered to be within the scope of protection of this invention.

Claims

1. A diagnostic electrocardiogram system, comprising: an electrode lead system (40) for generating at least one electrode signal indicative of electrical activity of a subject heart (10); and a diagnostic electrocardiograph (50) operably coupled to the electrode lead system (40) to communicate a subject electrocardiogram (20) and at least one diagnosed electrocardiogram (30) designated as morphologically matching the subject electrocardiogram (20), wherein the subject electrocardiogram (20) provides at least one interpreted information of ECG features derived from the electrical activity of the subject heart (10) as indicated by the at least one electrode signal, wherein the at least one diagnosed electrocardiogram (30) is accessible by the diagnostic electrocardiograph (50) and provides at least one diagnosed information of ECG features derived from electrical activity of a recorded at least one diagnosed heart (11), and wherein the designation by the diagnostic electrocardiograph (50) of the at least one diagnosed electrocardiogram (30) as morphologically matching the subject electrocardiogram (20) comprises: the diagnostic electrocardiograph (50) is further configured to determine a probability that the at least one diagnosed electrocardiogram (30) represents an accurate diagnostic assessment of the subject electrocardiogram (20) based on a similarity between a morphology of the subject electrocardiogram (20) and a morphology of the at least one diagnosed electrocardiogram (30), wherein the diagnostic electrocardiograph (50) comprises a subject electrocardiogram controller (61) and a diagnostic electrocardiogram controller (62); wherein, responsive to the generation of the at least one electrode signal by the electrode lead system (40), the subject electrocardiogram controller (61) is configured to control the generation of the subject electrocardiogram (20); and wherein, responsive to the generation of the subject electrocardiogram (20) by the subject electrocardiogram controller (61), the diagnostic electrocardiogram controller (62) is configured to control the designation of the at least one diagnosed electrocardiogram (30) as morphologically matching the subject electrocardiogram (20), wherein the control by the diagnostic electrocardiogram controller (62) of the designation of the at least one diagnosed electrocardiogram (30) as morphologically matching the subject electrocardiogram (20) comprises: the diagnostic electrocardiogram controller (62) is further configured to access a cluster tree constructed from a training set of diagnosed electrocardiograms (30) providing a plurality of diagnosed information of ECG features derived from electrical activity of a recorded plurality of diagnosed hearts (11); and the diagnostic electrocardiogram controller (62) is further configured to control navigation of the cluster tree to designate the at least one diagnosed electrocardiogram (30) from the training set of diagnosed electrocardiograms (30) as morphologically matching the subject electrocardiogram (20), wherein the navigation of the cluster tree by the diagnostic electrocardiogram controller (62) comprises: The diagnostic ECG controller (62) is also configured to generate a cheap ECG feature vector by applying a diagnostic ECG contour vector to the object ECG (20); and The diagnostic ECG controller (62) is also configured to derive a nearest neighbor list of at least one diagnostic ECG based on navigation along the clustering tree according to the cheap ECG feature vector. The navigation of the clustering tree by the diagnostic electrocardiogram controller (62) further includes: The diagnostic ECG controller (62) is further configured to calculate at least one expensive ECG feature between the target ECG (20) and the nearest neighbor list of at least one diagnostic ECG; and The diagnostic ECG controller is also configured to derive the ECG (30) of the at least one diagnosis based on the at least one expensive ECG feature as the designation that matches the morphology of the target ECG (20).

2. The diagnostic electrocardiogram system according to claim 1, wherein At least one interpretation of the ECG features derived from the electrical activity of the object heart (10) indicated by the at least one electrode signal includes at least one of the following: algorithmic interpretation and electrocardiogram operator interpretation; and The at least one diagnosis derived from the ECG features of the heart (11) based on the recorded electrical activity of at least one diagnosis includes at least one of the following: algorithmic diagnosis and electrocardiogram personnel diagnosis.

3. The diagnostic electrocardiograph system of claim 1, wherein, The control by the diagnostic electrocardiogram controller (62) for the at least one diagnostic electrocardiogram (30) as a morphological match with the specified electrocardiogram (20) includes: The diagnostic electrocardiogram controller (62) is also configured to control the allocation of the electrocardiogram (30) of the at least one diagnosis to at least one diagnostic category representing at least one diagnostic evaluation of the object electrocardiogram (20).

4. The diagnostic electrocardiograph system of claim 3, wherein, The control by the diagnostic electrocardiogram controller (62) for the at least one diagnostic electrocardiogram (30) as a morphological match with the specified electrocardiogram (20) includes: For multiple diagnostic categories, the diagnostic electrocardiogram controller (62) is also configured to control the determination of the accurate diagnostic probability for each diagnostic category.

5. A diagnostic electrocardiograph (50), comprising: Object electrocardiogram controller (61), In response to at least one electrode signal indicating electrical activity of the subject's heart (10), the subject electrocardiogram controller (61) is configured to control the generation of a subject electrocardiogram (20), the subject electrocardiogram (20) providing information on at least one interpretation of ECG features derived from the electrical activity of the subject's heart (10) indicated by the at least one electrode signal; and Diagnostic electrocardiogram controller (62), In response to the generation of the object electrocardiogram (20) by the object electrocardiogram controller (61), the diagnostic electrocardiogram controller (62) is configured to control the communication of at least one diagnostic electrocardiogram (30) as a specified communication that matches the morphology of the object electrocardiogram (20). The at least one diagnostic electrocardiogram (30) provides information on at least one diagnosis from ECG features derived from the electrical activity of the heart (11) recorded for at least one diagnosis, and The designation of the at least one diagnostic electrocardiogram (30) by the diagnostic electrocardiogram controller (62) as a morphological match with the target electrocardiogram (20) includes: The diagnostic electrocardiogram controller (62) is further configured to determine the probability that the at least one diagnostic electrocardiogram (30) represents an accurate diagnostic evaluation of the object electrocardiogram (20) based on the similarity between the morphology of the object electrocardiogram (20) and the morphology of the at least one diagnostic electrocardiogram (30). The control by the diagnostic electrocardiogram controller (62) for the at least one diagnostic electrocardiogram (30) as a morphological match with the target electrocardiogram (20) includes: The diagnostic electrocardiogram controller (62) is also configured to access a clustering tree constructed based on a training set of diagnostic electrocardiograms (30), which provide information on multiple diagnoses derived from ECG features of the recorded electrical activity of the heart (11) of multiple diagnoses; and The diagnostic electrocardiogram controller (62) is also configured to control the navigation of the clustering tree to designate the at least one diagnostic electrocardiogram (30) as morphologically matched with the object electrocardiogram (20). The navigation of the clustering tree by the diagnostic electrocardiogram controller (62) includes: The diagnostic ECG controller (62) is also configured to generate a cheap ECG feature vector by applying a diagnostic ECG contour vector to the object ECG (20); and The diagnostic ECG controller (62) is also configured to derive a nearest neighbor list of at least one diagnostic ECG based on navigation along the clustering tree according to the cheap ECG feature vector. The navigation of the clustering tree by the diagnostic electrocardiogram controller (62) further includes: The diagnostic ECG controller (62) is further configured to calculate at least one expensive ECG feature between the target ECG (20) and the nearest neighbor list of at least one diagnostic ECG; and The diagnostic ECG controller is also configured to derive the ECG (30) of the at least one diagnosis based on the at least one expensive ECG feature as the designation that matches the morphology of the target ECG (20).

6. The diagnostic electrocardiograph (50) according to claim 5, wherein At least one interpretation of the ECG features derived from the electrical activity of the object heart (10) indicated by the at least one electrode signal includes at least one of the following: algorithmic interpretation and electrocardiogram operator interpretation; and The at least one diagnosis derived from the recorded electrical activity of the heart (11) of at least one diagnosis includes at least one of the following: algorithmic diagnosis and electrocardiogram personnel diagnosis.

7. The diagnostic electrocardiograph (50) according to claim 5, wherein, The control by the diagnostic electrocardiogram controller (62) for the at least one diagnostic electrocardiogram (30) as a morphological match with the specified electrocardiogram (20) includes: The diagnostic electrocardiogram controller (62) is also configured to control the allocation of the electrocardiogram (30) of the at least one diagnosis to at least one diagnostic category representing at least one diagnostic evaluation of the object electrocardiogram (20).

8. The diagnostic electrocardiograph (50) according to claim 7, wherein, The control by the diagnostic electrocardiogram controller (62) for the at least one diagnostic electrocardiogram (30) as a morphological match with the specified electrocardiogram (20) includes: For multiple diagnostic categories, the diagnostic electrocardiogram controller (62) is configured to control the determination of the accurate diagnostic probability for each diagnostic category.

Images