Ladder gram generating device and computer program

Abstract

To provide a technology capable of generating a laddergram. The laddergram generation device includes an acquisition unit that acquires a specific electrocardiogram of a subject, and a calculation unit. The calculation unit generates a laddergram from the specific electrocardiogram based on machine learning using training data including a training electrocardiogram and a training laddergram corresponding to the training electrocardiogram.

Description

[Technical Field]

[0001] The technology disclosed in this specification relates to a technology for generating a laddergram from an electrocardiogram. [Background technology]

[0002] Patent Document 1 discloses a Holter electrocardiograph used to measure electrocardiograms. When diagnosing cardiac disease (e.g., arrhythmia) using an electrocardiogram, a laddergram may be created from the electrocardiogram to consider the complex behavior of the cardiac conduction system. A laddergram is a diagram that shows how electrical signals are transmitted to each part of the heart. Creating a laddergram makes it easier to understand abnormalities in excitation conduction within the heart. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-088643 Summary of the Invention [Problem to be solved by the invention]

[0004] Since laddergrams are created manually by medical professionals such as doctors, it takes time for a doctor to determine whether an electrocardiogram is taken or not, such as cardiac arrhythmia. This specification discloses a technology for automatically generating laddergrams. [Means for solving the problem]

[0005] In a first aspect of the technology disclosed in this specification, a laddergram generation device includes an acquisition unit that acquires a specific electrocardiogram of a subject and a calculation unit. The calculation unit generates a laddergram from the specific electrocardiogram based on machine learning using training data including a training electrocardiogram and a training laddergram corresponding to the training electrocardiogram.

[0006] In the laddergram generation device, the calculation unit automatically generates a laddergram corresponding to a specific electrocardiogram based on machine learning using a training laddergram, eliminating the need for medical professionals to generate laddergrams themselves.

[0007] This specification also discloses a computer program for generating a laddergram, which causes a computer to function as a generator that generates a laddergram from a specific electrocardiogram of a subject based on machine learning using training data including a training electrocardiogram and a training laddergram corresponding to the training electrocardiogram. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1(a) is a diagram showing the structure of the heart, and FIG. 1(b) is a diagram showing the relationship between an electrocardiogram and a laddergram. [Figure 2] FIG. 1 is a block diagram showing the functions of a ladder diagram generation system according to an embodiment. [Figure 3] 10 is a flowchart showing an example of machine learning processing. [Figure 4] FIG. 1 is a diagram showing the relationship between an electrocardiogram and a segment laddergram. [Figure 5] (a) shows an electrocardiogram of a case of second-degree atrioventricular block, and (b) shows a segmented laddergram corresponding to (a). [Figure 6] 1 is a flowchart showing an example of a generation process for generating a laddergram from a subject's electrocardiogram using a deep neural network (DNN) trained by machine learning processing. [Figure 7] (a) shows the electrocardiogram of a normal person (normal case), and (b) shows the segmented laddergram corresponding to (a) and arrows indicating the direction of electrical signal (action potential) propagation. [Figure 8] (a) shows an electrocardiogram of an example of atrioventricular block without escape contractions or escape rhythm from the atrioventricular junction or ventricles, and (b) shows a segmented laddergram corresponding to (a) and arrows indicating the direction of electrical signal transmission. [Figure 9](a) shows an electrocardiogram of an example of atrioventricular block with escape beats (or escape rhythm) from the ventricles, and (b) shows a segmented laddergram corresponding to (a) and arrows indicating the direction of electrical signal transmission. [Figure 10] (a) shows an electrocardiogram of an example of atrioventricular block with escape beats (or escape rhythm) from the atrioventricular junction, and (b) shows a segmented laddergram corresponding to (a) and arrows indicating the direction of electrical signal transmission. [Figure 11] 10A and 10B are diagrams for explaining a method for creating a line laddergram, in which (a) shows an electrocardiogram, (b) shows a segmented laddergram corresponding to (a), and (c) shows a line laddergram generated based on (b). DETAILED DESCRIPTION OF THE INVENTION

[0009] The main features of the embodiments described below are listed below. Note that the technical elements described below are independent technical elements that exhibit technical utility alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing.

[0010] In a second aspect of the technology disclosed in this specification, in the first aspect described above, the training ladder gram may be configured with time information related to times representing the onset and completion of excitation of each of multiple sites within the heart. The calculation unit may generate a ladder gram based on machine learning, the ladder gram being configured with specific time information related to specific times representing the onset and completion of excitation of each of multiple sites within the heart in a specific electrocardiogram. With this configuration, it is possible to generate a ladder gram that quantitatively shows the transmission of electrical signals at each site within the heart.

[0011] In a third aspect of the technology disclosed herein, in the second aspect described above, the training ladder gram may be drawn using segments separated at specific times based on specific time information. Furthermore, in a fourth aspect of the technology disclosed herein, in the third aspect described above, the ladder gram (i.e., the generated ladder gram) may be drawn using segments separated at specific times based on specific time information. By drawing the training ladder gram or a ladder gram generated using a ladder gram generation device using segments as described above, the basis for drawing the ladder gram can be clarified.

[0012] In a fifth aspect of the technology disclosed in the present specification, in any one of the second to fourth aspects, the calculation unit may identify the direction of transmission of electrical signals between multiple parts in a specific electrocardiogram based on the specific time information. With this configuration, a medical professional such as a doctor who is an examiner (hereinafter simply referred to as a medical professional) can easily understand the direction of transmission of electrical signals between multiple parts in the heart.

[0013] In a sixth aspect of the technology disclosed in this specification, the laddergram generation device in the fourth aspect may further include an output unit that outputs the laddergram generated by the calculation unit. The output unit may output a laddergram drawn using segments. With this configuration, a laddergram composed of specific time information is drawn using segments separated by specific times. This makes it easier for medical professionals to visually grasp the times at which each part of the heart is excited.

[0014] In a seventh aspect of the technology disclosed herein, in the sixth aspect described above, the output unit may further output a line diagram laddergram that shows the laddergram in a line diagram. The line diagram laddergram may be a graph in which multiple sites are shown on the vertical axis in accordance with the order of transmission of electrical signals (action potentials) and time is shown on the horizontal axis. Each of the multiple sites in the graph may have an upper end and a lower end. A line segment may connect the estimated start time of the site on one of the upper end and the lower end with the estimated end time of the site on the other of the upper end and the lower end. Conventional laddergrams manually created by medical professionals are shown in line diagrams. By outputting a line diagram laddergram, it is possible to output a line diagram laddergram that is traditionally familiar to medical professionals. This makes it easier for medical professionals to grasp the overall abnormality in electrical signal transmission between each site in the heart.

[0015] In an eighth aspect of the technology disclosed in this specification, in any one of the first to seventh aspects, the calculation unit may identify a candidate for heart disease based on the generated ladder diagram. With this configuration, it is possible to provide a doctor with information to assist in diagnosis. [Example]

[0016] A laddergram generation system 1 according to an embodiment will be described with reference to the drawings. The laddergram generation system 1 generates a laddergram from a subject's electrocardiogram. As shown in FIG. 1(a), electrical excitation that causes heartbeats is generated in the sinoatrial node (SA) and transmitted through the atria (A), atrioventricular node (AV), and ventricles (V) in that order. A laddergram is a diagram illustrating electrical propagation between the sinoatrial node (SA), atria (A), atrioventricular node (AV), and ventricles (V). FIG. 1(b) shows a portion of an electrocardiogram and a laddergram corresponding to the electrocardiogram. In FIG. 1(b), the laddergram is shown as a line diagram. Generally, conventional laddergrams manually created by medical professionals such as doctors (hereinafter simply referred to as medical professionals) are created as line diagrams. Creating a laddergram organizes the conduction of stimuli between the atria (A) and ventricles (V), allowing for a quick understanding of excitation propagation in the stimulus conduction system. In this specification, the term "electrocardiogram" may refer to either the waveform itself, which represents the electrical signal of the cardiac conduction system of the subject, or data representing the waveform. Furthermore, the laddergram is not limited to a laddergram shown as a line diagram. In this embodiment, as will be described in detail below, the laddergram includes not only laddergrams shown as line diagrams, but also laddergrams shown in other forms (e.g., segmented laddergrams, etc.). A laddergram can also be considered a diagram that represents electrical propagation to each site in the heart in a time series. In this embodiment, candidates for arrhythmia diagnosis are identified from laddergrams generated by the laddergram generation system 1, as will be described later, but the present invention is not limited to such a configuration. Candidates for cardiac diseases other than arrhythmia may also be identified as long as they can be identified from the generated laddergram.

[0017] As shown in FIG. 2, the ladder gram generation system 1 includes a machine learning device 10 and a ladder gram generation device 30.

[0018] The machine learning device 10 includes a calculation unit 12 and a communication unit 24. The calculation unit 12 is connected to the communication unit 24 via wiring such as a bus bar so that they can communicate with each other.

[0019] The calculation unit 12 is configured using a computer including a memory 14 and a CPU 20. The memory 14 includes hardware such as a ROM and a RAM. A calculation program is stored in the memory 14, and when the CPU 20 executes the calculation program, the CPU 20 functions as a learning processing unit 22 shown in FIG. 2. The processing of the learning processing unit 22 will be described in detail later. The memory 14 includes a learning data storage unit 16 and a trained DNN storage unit 18.

[0020] The learning data storage unit 16 stores data used in machine learning processing (hereinafter also referred to as learning data). The learning data is used in machine learning processing to generate a ladder diagram, which will be described later. The learning data is acquired from an external device via the communication unit 24 and stored in the learning data storage unit 16.

[0021] The trained DNN storage unit 18 stores information including a machine learning model trained by machine learning. In this embodiment, features for generating a laddergram from an electrocardiogram are learned by machine learning using a deep neural network (DNN). The trained DNN storage unit 18 stores a DNN trained by machine learning (hereinafter also referred to as a trained DNN).

[0022] The ladder diagram generation device 30 includes a calculation unit 32, a communication unit 44, and an output unit 46. The calculation unit 32 is connected to the communication unit 44 and the output unit 46 via wiring such as a bus bar so that they can communicate with each other.

[0023] The calculation unit 32 is configured using a computer including a memory 34 and a CPU 40. The memory 34 includes hardware such as a ROM and a RAM. A calculation program is stored in the memory 34, and when the CPU 40 executes the calculation program, the CPU 40 functions as a generation processing unit 42 shown in FIG. 2. The processing of the generation processing unit 42 will be described in detail later. The memory 34 includes a subject information storage unit 36 ​​and a trained DNN storage unit 38.

[0024] The communication unit 24 includes a communication interface that enables the machine learning device 10 to communicate with an external device via a wired or wireless connection. The calculation unit 12 can acquire various types of data from the external device via the communication unit 24. Specifically, the calculation unit 12 acquires training data transmitted from the external device via the communication unit 24. The calculation unit 12 stores the acquired training data in the training data storage unit 16.

[0025] The subject information storage unit 36 ​​stores information about the subject. The information about the subject includes the subject's electrocardiogram (hereinafter also referred to as a specific electrocardiogram), a laddergram of the specific electrocardiogram generated by the generation processing unit 42, and arrhythmia diagnosis candidates identified from the generated laddergram. The subject's electrocardiogram (specific electrocardiogram) is measured using an electrocardiogram measuring device such as a Holter electrocardiograph. The calculation unit 32 acquires the specific electrocardiogram from an external device such as an electrocardiogram measuring device via the communication unit 44 and stores it in the subject information storage unit 36. The calculation unit 32 stores the laddergram of the specific electrocardiogram generated by the generation processing unit 42 and the identified arrhythmia diagnosis candidates in the subject information storage unit 36.

[0026] The trained DNN storage unit 38 stores trained DNNs trained by machine learning in the machine learning device 10. The calculation unit 32 acquires the trained DNNs from the machine learning device 10 via the communication units 24 and 44, and stores them in the trained DNN storage unit 38.

[0027] The memory 14 also stores diagnostic information used in the process of identifying arrhythmia diagnosis candidates (described later), ladder gram corresponding to the diagnostic information, and information derived therefrom (hereinafter also referred to as information on arrhythmia diagnosis candidates). The diagnostic information includes ladder gram patterns that are suspected of arrhythmia. The memory 14 associates ladder gram patterns that are suspected of arrhythmia with the type of diagnosis (i.e., disease name) and stores them as information on arrhythmia diagnosis candidates.

[0028] The communication unit 44 includes a communication interface for the laddergram generation device 30 to communicate with an external device via a wired or wireless connection. The calculation unit 32 can acquire various data from the external device via the communication unit 44. Specifically, the calculation unit 32 acquires, via the communication unit 44, an electrocardiogram (specific electrocardiogram) of the subject transmitted from the external device and a trained DNN transmitted from the external device (specifically, the machine learning device 10). The calculation unit 32 stores the acquired specific electrocardiogram in the subject information storage unit 36 ​​and stores the acquired trained DNN in the trained DNN storage unit 38.

[0029] The output unit 46 outputs the laddergram generated by the calculation unit 32 and the identified arrhythmia candidates. In this embodiment, the output unit 46 is a display device, and presents the generated laddergram and the identified arrhythmia candidates to the medical professional by displaying the generated laddergram and the identified arrhythmia candidates on the display device. The output unit 46 may be a printer, and may present the generated laddergram and the identified arrhythmia candidates to the medical professional by printing the generated laddergram and the identified disease candidates on the printer. The laddergram generation device 30 may also include both a monitor and a printer as the output unit 46, or may be connectable to an external display device and printer.

[0030] Next, we will explain the process of generating a laddergram from the electrocardiogram of a subject by the laddergram generation system 1. The process of generating a laddergram from the electrocardiogram of a subject by the laddergram generation system 1 is executed by a machine learning process in which the machine learning device 10 learns, by machine learning, features for generating a laddergram from the electrocardiogram, and a generation process in which the laddergram generation device 30 generates a laddergram from the electrocardiogram of the subject using a trained DNN learned by the machine learning process.

[0031] First, the machine learning process executed by the machine learning device 10 will be described. As shown in FIG. 3, first, the learning processing unit 22 acquires learning data (S100). The learning data is stored in the learning data storage unit 16. The learning processing unit 22 acquires the learning data stored in the learning data storage unit 16. The learning data is data for learning features for generating a laddergram from an electrocardiogram, and includes training electrocardiogram data and its correct answer labels. The training data includes supervised data. The correct answer labels are laddergrams manually generated by a specialist from the training electrocardiogram data. The learning processing unit 22 performs DNN machine learning using the training data. Note that known DNN machine learning techniques can be applied to DNN machine learning, so a detailed description will be omitted.

[0032] Here, the ladder gram used as the correct label will be described. As shown in FIG. 4, in this embodiment, the ladder gram used as the correct label is a ladder gram (hereinafter also referred to as a segmented ladder gram) that shows the transmission of electrical signals (action potentials) at each site in the heart using four-layered segments. In the segmented ladder gram of FIG. 4, the vertical axis indicates each site in the heart, and the horizontal axis indicates time. On the vertical axis, each site is arranged from top to bottom according to the order of normal excitation conduction to each site in the heart. In this embodiment, the sites arranged on the vertical axis are, from top to bottom, the sinoatrial node (SA), atrium (A), atrioventricular junction (AVJ), and ventricle (V). A segment indicates the time from excitation start to excitation completion of the corresponding site. Specifically, the excitation start time of each site is the time when it is estimated that one of the cells in that site is excited. Furthermore, the excitation completion time of each site is the time when it is estimated that excitation has spread to all the cells in that site. That is, a segment indicates, as quantitative time information, the time from when excitation reaches a corresponding region to when the excitation spreads throughout the entire corresponding region.

[0033] In a segmented laddergram, the end of a segment on the excitation completion time side may overlap with the end of the excitation start time side of the segment of the next region (the adjacent region below on the vertical axis). For example, in FIG. 4, a portion of the right side of the segment representing the atrium (A) overlaps in time with a portion of the left side of the atrioventricular junction (AVJ). This indicates that an electrical signal is transmitted from the atrium (A) to the atrioventricular junction (AVJ) before the entire atrium (A) is excited. Furthermore, the end of a segment on the excitation completion time side may substantially coincide with the end of the excitation start time side of the segment of the next region (the adjacent region below on the vertical axis). For example, in FIG. 4, the right end of the segment representing the atrioventricular junction (AVJ) substantially coincides with the left end of the segment representing the ventricle (V). This indicates that excitation of the ventricle (V) begins substantially simultaneously with the completion of excitation of the atrioventricular junction (AVJ).

[0034] Next, the learning processing unit 22 executes machine learning processing using the learning data acquired in step S100 (S110). The learning data includes an electrocardiogram and a segmented laddergram corresponding to the electrocardiogram. In the machine learning processing, the features of each segment of the segmented laddergram and the electrocardiogram corresponding to the segment are learned. The features learned here include not only the shape of the electrocardiogram waveform from the excitation start time to the excitation end time of a specific segment, but also the features of the electrocardiogram waveform over a continuous time interval including multiple segments.

[0035] The training data used include not only electrocardiograms of people with normal rhythm (normal sinus rhythm) without arrhythmia (hereinafter simply referred to as "normal people"), but also electrocardiograms of patients with various arrhythmias. For example, FIG. 5(a) shows an electrocardiogram of a patient with atrioventricular block, and FIG. 5(b) shows a segmented laddergram corresponding to FIG. 5(a). As shown in FIGS. 5(a) and 5(b), in a patient with atrioventricular block, there are a mixture of heartbeats in which excitation is transmitted along the cardiac conduction system in the order of the sinoatrial node (SA), atrium (A), atrioventricular junction (AVJ), and ventricle (V), and heartbeats in which excitation is transmitted to the sinoatrial node (SA), atrium (A), and atrioventricular junction (AVJ) and then discontinued within the atrioventricular junction (AVJ). Note that because excitation is not transmitted across the entire atrioventricular junction (AVJ), the segment representing the electrical signal transmitted to the AVJ is omitted in the segmented laddergram shown in Figure 5(b).By conducting machine learning using the electrocardiograms of patients with arrhythmias such as atrioventricular block, it is possible to learn the characteristics of when electrical signals are not transmitted normally to each part of the heart.

[0036] When the machine learning process of step S110 is completed, the learning processing unit 22 stores the DNN learned in step S110 in the learned DNN storage unit 18 (S120).

[0037] Next, a generation process for generating a laddergram from a subject's electrocardiogram using a trained DNN executed by the laddergram generation device 30 will be described. As shown in FIG. 6, first, the generation processing unit 42 acquires a specific electrocardiogram stored in the subject information storage unit 36 ​​(S200). As described above, the specific electrocardiogram is an electrocardiogram of the subject measured using an electrocardiogram measuring device such as a Holter electrocardiograph. The specific electrocardiogram is stored in the subject information storage unit 36. The generation processing unit 42 acquires the specific electrocardiogram from the subject information storage unit 36. Next, the generation processing unit 42 acquires the trained DNN stored in the trained DNN storage unit 18 (S210).

[0038] Next, the generation processing unit 42 generates a segment ladder gram from the specific electrocardiogram acquired in step S200 using the trained DNN acquired in step S210 (S220). The trained DNN acquired in step S210 has learned features for generating segments of each part in the heart from the electrocardiogram. Therefore, by using the trained DNN, it is possible to identify segments corresponding to the specific electrocardiogram and generate a segment ladder gram from the identified segments.

[0039] Next, the generation processing unit 42 determines the direction of electrical signal transmission between each segment of the segmented laddergram generated in step S220 (S230). Assuming a proximal end (SA side) and a distal end (V side) of the cardiac conduction system, the direction of transmission from the proximal end to the distal end is defined as forward conduction, and conversely, the direction of transmission from the distal end to the proximal end is defined as retrograde conduction. Once excited, cardiomyocytes constituting the heart enter a refractory period and cannot enter the next excitation until the refractory period due to the preceding excitation disappears. Specifically, the generation processing unit 42 determines that an electrical signal has been transmitted between adjacent regions on the vertical axis of the segmented laddergram if the adjacent regions overlap or if the end point and the start point are approximately the same. The generation processing unit 42 then determines that an electrical signal has been transmitted from the first excited region to the second excited region.

[0040] Figure 7(b) shows a segmented laddergram of a normal person. In Figure 7(b), the direction of electrical signal transmission is indicated by arrows. In a normal person, electrical signals are transmitted in order from the top to the bottom of the segmented laddergram.

[0041] Figure 8(b) shows a segmented laddergram of a patient with atrioventricular block without ectopic spontaneous excitation (no escape beats or rhythms due to spontaneous excitation from sites other than the sinoatrial node (SA)). The heartbeat on the left side of Figure 8(b) is a blocked beat, in which an electrical signal is transmitted in the forward direction from the upper to the lower chambers to the sinoatrial node (SA) and atrium (A), and excitation conduction is terminated within the atrioventricular junction (AVJ). The segmented laddergram shown in Figure 8(b) omits the electrical signal that has been transmitted partway through the atrioventricular junction (AVJ). On the other hand, the heartbeat on the right side of Figure 8(b) is a normal beat, in which an electrical signal is transmitted in the forward direction from the sinoatrial node (SA), atrium (A), and atrioventricular junction (AVJ), but is not terminated within the atrioventricular junction (AVJ), but is transmitted throughout the atrioventricular junction (AVJ), and then further transmitted in the forward direction from the atrioventricular junction (AVJ) to the ventricle (V).

[0042] Figure 9(b) shows a segmented laddergram of a patient with atrioventricular block (AVB) and ectopic spontaneous excitation (escape beats and rhythms due to spontaneous excitation from sites other than the SA node). In the two heartbeats shown in Figure 9(b), forward electrical signals are transmitted to the SA node and atrium (A), but excitation conduction is blocked within the AVJ. Corresponding to the AVB, reverse electrical signals originating from the ventricle (V) and traveling from the lower to the upper chambers are also observed, which are blocked within the AVJ. Note that neither the electrical signals transmitted from the atrium (A) to the AVJ nor those transmitted from the ventricle (V) to the AVJ propagate throughout the AVJ. Therefore, the segment representing the electrical signals transmitted to the AVJ is omitted from the segmented laddergram shown in Figure 9(b). The most reasonable explanation for this atrial (A) and ventricular (V) activation pattern is ventricular escape rhythm secondary to atrioventricular block.

[0043] Figure 10(b) is a segmental ladder gram of a different case from Figure 9(b), showing atrioventricular block with ectopic spontaneous excitation. The source of excitation is different from that of Figure 9(b). During the 3-second period shown in Figure 10(b), excitation from the sinoatrial node (SA) (at a rate of 92–93 beats per minute) is transmitted forward from the SA to the atrium (A) and terminates within the atrioventricular junction (AVJ). The atrioventricular junction (AVJ) is excited a short time after excitation halfway through the AVJ. This is presumably due to the generation of ectopic spontaneous excitation (escape contractions / escape rhythm) at a rate of 44 beats per minute from the AVJ in response to high-degree atrioventricular block. This ectopic spontaneous excitation generated at the AVJ is conducted forward from the AVJ to the ventricle (V), but is blocked by the refractory period and does not conduct backward. In the segmented laddergram shown in Figure 10(b), the electrical signals transmitted from the atrium (A) to partway through the atrioventricular junction (AVJ) are omitted, and only the ectopic spontaneous excitation occurring at the atrioventricular junction (AVJ) and the downstream ventricular (V) excitation are depicted.

[0044] In Figures 7(b), 8(b), 9(b), and 10(b), the direction of electrical signal transmission between each segment is identified and indicated by arrows. Identifying the direction of electrical signal transmission between each segment in this way makes it easier to understand the state of excitation conduction in each part of the heart, which is useful for electrophysiological diagnosis of arrhythmia.

[0045] Next, the generation processing unit 42 generates a line laddergram based on the segment laddergram generated in step S220 and the direction of electrical signal transmission between the segments identified in step S230 (S240). Specifically, the generation processing unit 42 identifies segments through which electrical signals are transmitted between adjacent sites in the vertical direction. The generation processing unit 42 then connects, with a line segment, a point representing the estimated time of the start of excitation at the upper or lower end of the segment through which the electrical signal is transmitted (hereinafter also referred to as the start point) to a point representing the estimated time of the end of excitation at the lower or upper end (hereinafter also referred to as the end point). In this case, if the electrical signal is transmitted in the forward direction, the start point is located on the upper end and the end point is located on the lower end. On the other hand, if the electrical signal is transmitted in the reverse direction, the start point is located on the lower end and the end point is located on the upper end. The generation processing unit 42 also identifies the site and time when the excitation occurred based on the segment through which the electrical signal is transmitted between adjacent sites in the vertical direction and the direction of transmission of the electrical signal between the segments, and marks a point indicating the excitation occurrence between the upper and lower ends of the identified site. The generation processing unit 42 connects the point indicating the excitation occurrence to the starting point of the site where the excitation occurred and the adjacent site.

[0046] Here, a method for creating a linear laddergram will be described in more detail with reference to FIG. 11. FIG. 11 illustrates a case where an electrical signal is transmitted in the forward direction. First, the generation processing unit 42 identifies the excitation onset of the atrium (A) (the origin of the P wave in an electrocardiogram). The atrium (A) is drawn so that the excitation onset and excitation completion times of the atrium (A) are the same. The excitation onset of the atrium (A) and the excitation onset of the atrioventricular junction (AVJ) are assumed to be the same time. As shown in FIG. 11(b), in a segmented laddergram, the excitation onset of the atrioventricular junction (AVJ) is slightly (approximately 16 ms) later than the excitation onset of the atrium (A), but in a linear laddergram, this delay is considered to be within a negligible range. Next, the generation processing unit 42 identifies the excitation completion time of the atrioventricular junction (AVJ). The excitation completion time of the atrioventricular junction (AVJ) coincides with the excitation start time of the ventricle (V) (the start time of the QRS wave in the electrocardiogram). Next, the generation processing unit 42 connects the excitation start time and excitation completion time of the atrioventricular junction (AVJ) with a straight line. This results in the atrioventricular junction (AV). Next, the generation processing unit 42 identifies the excitation completion time of the ventricle (V) (the end time of the QRS wave in the electrocardiogram) and connects the excitation start time of the ventricle (V) (the excitation completion time of the atrioventricular junction (AVJ)) with the excitation completion time of the ventricle (V) with a straight line. Alternatively, the generation processing unit 42 may assume an excitation source in the middle (between the upper and lower ends) of the sinoatrial node (SA), and may transmit excitation from the sinoatrial node (SA) to the atrium (A) over approximately 40 ms. Next, a straight line is connected from the excitation source of the sinoatrial node (SA) to the excitation time of the atrium (A). As a result, the ladder diagram shown in FIG. 11(c) is created.

[0047] The generation processing unit 42 also identifies a presumed arrhythmia diagnosis candidate (hereinafter also referred to as a disease candidate) based on the segment ladder gram generated in step S220 and the inter-segment electrical propagation direction identified in step S230 (S250). Specifically, the generation processing unit 42 reads information related to the arrhythmia diagnosis candidate from the memory 14. The generation processing unit 42 then determines whether the generated segment ladder gram includes a segment ladder gram and a pattern of inter-segment electrical signal propagation direction included in the information related to the arrhythmia diagnosis candidate. If the generated segment ladder gram includes a ladder gram pattern included in the information related to the arrhythmia diagnosis candidate, the generation processing unit 42 identifies the type of arrhythmia (disease name) associated with the ladder gram pattern as a disease candidate. This makes it possible to provide a doctor with information to assist in diagnosis. For example, as shown in Figure 8(b), if an electrical signal is transmitted to the sinoatrial node (SA) and atrium (A) but not to the atrioventricular junction (AVJ) in a segmented laddergram, it can be inferred that there is atrioventricular block without ectopic spontaneous excitation.Also, as shown in Figure 9(b), if an electrical signal is transmitted to the sinoatrial node (SA) and atrium (A) but not to the atrioventricular junction (AVJ) in a segmented laddergram, it can be inferred that there is atrioventricular block with escape contractions (or escape rhythm) from the ventricles. 10(b), when an electrical signal is transmitted to the sinoatrial node (SA) and the atrium (A) but not to the atrioventricular junction (AVJ), and an electrical signal is generated at the atrioventricular junction (AVJ) and transmitted from the atrioventricular junction (AVJ) to the ventricle (V), it can be estimated that there is an atrioventricular block with escape contractions (or escape rhythm) from the atrioventricular junction. The generation processing unit 42 may also identify a disease candidate from the diagram laddergram generated in step S240. In this embodiment, after generating the diagram laddergram in step S240, an arrhythmia diagnosis candidate is identified in step S250. However, the present invention is not limited to such a configuration.For example, after identifying a candidate arrhythmia diagnosis in step S250, a diagrammatic laddergram may be generated in step S240, or the generation of the diagrammatic laddergram in step S240 and the identification of a candidate arrhythmia diagnosis in step S250 may be performed simultaneously.

[0048] Next, the generation processing unit 42 outputs the diagram laddergram generated in step S240 and, if a disease candidate is identified in step S250, the identified disease candidate (S260). For example, if the output unit 46 is a display device, the generation processing unit 42 displays the diagram laddergram and the disease candidate on the display device. If the output unit 46 is a printer, the generation processing unit 42 prints and outputs the diagram laddergram and the disease candidate. As described above, diagram laddergrams have traditionally been created manually by medical professionals. By outputting the diagram laddergram, a laddergram in a format familiar to medical professionals can be presented to the medical professionals. This makes it easier for medical professionals to understand the excitation propagation of the stimulus conduction system of a specific electrocardiogram.

[0049] The generation processing unit 42 may output, together with the line ladder gram, a ladder gram (hereinafter also referred to as a digital ladder gram) consisting of the segment ladder gram generated in step S220 and marks (e.g., arrows) indicating the transmission direction of electrical signals between the segments identified in step S230, or may output the segment ladder gram generated in step S220 (i.e., a segment ladder gram that does not include marks indicating the transmission direction of electrical signals between the segments).The generation processing unit 42 may also output only a digital ladder gram or only a segment ladder gram instead of the line ladder gram.

[0050] In addition, in this embodiment, the segment laddergram and the line laddergram have four layers (sinoatrial node (SA), atrium (A), atrioventricular junction (AVJ), and ventricle (V)), but are not limited to this configuration. For example, the segment laddergram and the line laddergram may have five or more layers. Furthermore, the layers of the segment laddergram and the line laddergram may be different. For example, the segment laddergram may be generated with five layers, and the layer of the line laddergram may be generated with four layers.

[0051] In this embodiment, machine learning is performed using an electrocardiogram and a segment laddergram as training data to generate a segment laddergram from a specific electrocardiogram, but this configuration is not limited to this. For example, machine learning may be performed using an electrocardiogram and a line laddergram as training data to generate a line laddergram from a specific electrocardiogram. In this case, too, the line laddergram is automatically generated by machine learning, allowing medical professionals to obtain a line laddergram from a specific electrocardiogram without spending time and effort.

[0052] However, as described in this embodiment, it is preferable to perform machine learning using an electrocardiogram and a segmented laddergram as learning data to generate a segmented laddergram from a specific electrocardiogram. As described above, the segmented laddergram can indicate, for each region, quantitative time information from the time when one of the cells in that region is estimated to be excited to the time when the excitation is estimated to have spread to all the cells in that region. Therefore, by evaluating the segmented laddergram, it is possible to determine the electrophysiological validity and consistency, and to more accurately understand the excitation propagation in the conduction system of the specific electrocardiogram.

[0053] Although specific examples of the technology disclosed in this specification have been described in detail above, these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and variations of the specific examples exemplified above. Furthermore, the technical elements described in this specification or drawings exhibit technical utility alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. Furthermore, the technology exemplified in this specification or drawings simultaneously achieves multiple objectives, and achieving one of those objectives itself has technical utility. [Explanation of symbols]

[0054] 1: Laddergram System 10: Machine learning device 12: Arithmetic section 14: Memory 16: Learning data storage unit 18: Trained DNN memory unit 20:CPU 22: Learning processing unit 24: Communications Department 30: Laddergram generator 32: Arithmetic section 34: Memory 36: Subject information storage unit 38: Trained DNN memory unit 40:CPU 42: Generation processing unit 44: Communications Department 46: Output section

Claims

1. an acquisition unit for acquiring a specific electrocardiogram of a subject; a calculation unit; and the calculation unit generates a ladder gram from the specific electrocardiogram based on machine learning using learning data including a learning electrocardiogram and a learning ladder gram corresponding to the learning electrocardiogram; the training laddergram is composed of time information relating to times representing the start and completion of excitation of each of a plurality of sites within the heart; The calculation unit generates a ladder gram based on the machine learning, the ladder gram being composed of specific time information related to specific times representing the start and completion of excitation of each of multiple sites within the heart in the specific electrocardiogram.

2. The ladder gram generating device according to claim 1 , wherein the training ladder gram is drawn using segments separated by the specific times based on the specific time information.

3. The laddergram generating device according to claim 2 , wherein the laddergram is drawn using segments separated by the specific times based on the specific time information.

4. The laddergram generating device according to claim 1 , wherein the calculation unit determines a direction of transmission of the electrical signal between the plurality of parts in the specific electrocardiogram based on the specific time information.

5. an output unit that outputs the ladder diagram generated by the calculation unit, The laddergram generating device according to claim 3 , wherein the output unit outputs the laddergram drawn using the segments.

6. the output unit further outputs a diagrammatic ladder diagram that shows the ladder diagram; The ladder diagram is a graph in which the vertical axis indicates the plurality of parts in the order of transmission of electrical signals and the horizontal axis indicates time, each of the plurality of portions of the graph having an upper end and a lower end; 6. The laddergram generating device according to claim 5, wherein the estimated start time of the part on one of the upper end and the lower end and the estimated completion time of the part on the other of the upper end and the lower end are connected by a line segment.

7. The laddergram generating device according to claim 1 , wherein the calculation unit identifies a candidate for heart disease based on the generated laddergram.

8. A computer program for generating a ladder diagram, comprising: Computer, a generating unit that generates a ladder gram from a specific electrocardiogram of a subject based on machine learning using learning data including a learning electrocardiogram and a learning ladder gram corresponding to the learning electrocardiogram; the training laddergram is composed of time information relating to times representing the start and completion of excitation of each of a plurality of sites within the heart; The generation unit generates a ladder gram in the specific electrocardiogram based on the machine learning, the ladder gram being composed of specific time information related to specific times representing the start and completion of excitation of each of multiple parts within the heart.

Images