WO2022244291A1 - プログラム、出力装置及びデータ処理方法 - Google Patents
プログラム、出力装置及びデータ処理方法 Download PDFInfo
- Publication number
- WO2022244291A1 WO2022244291A1 PCT/JP2021/047643 JP2021047643W WO2022244291A1 WO 2022244291 A1 WO2022244291 A1 WO 2022244291A1 JP 2021047643 W JP2021047643 W JP 2021047643W WO 2022244291 A1 WO2022244291 A1 WO 2022244291A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- electrocardiogram data
- fibrillation
- clear
- wave
- waves
- Prior art date
Links
- 238000003672 processing method Methods 0.000 title claims description 7
- 206010061592 cardiac fibrillation Diseases 0.000 claims abstract description 180
- 230000002600 fibrillogenic effect Effects 0.000 claims abstract description 171
- 238000010801 machine learning Methods 0.000 claims abstract description 74
- 230000000694 effects Effects 0.000 claims abstract description 13
- 230000036982 action potential Effects 0.000 claims abstract description 11
- 210000004165 myocardium Anatomy 0.000 claims abstract description 11
- 230000008859 change Effects 0.000 claims description 7
- 230000002123 temporal effect Effects 0.000 abstract 1
- 206010003658 Atrial Fibrillation Diseases 0.000 description 13
- 238000012545 processing Methods 0.000 description 11
- 238000000034 method Methods 0.000 description 10
- 238000004891 communication Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 9
- 238000013527 convolutional neural network Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 208000019622 heart disease Diseases 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
- A61B5/349—Detecting specific parameters of the electrocardiograph cycle
- A61B5/361—Detecting fibrillation
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
- A61B5/347—Detecting the frequency distribution of signals
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
- A61B5/349—Detecting specific parameters of the electrocardiograph cycle
- A61B5/352—Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
- A61B5/349—Detecting specific parameters of the electrocardiograph cycle
- A61B5/36—Detecting PQ interval, PR interval or QT interval
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/74—Details of notification to user or communication with user or patient ; user input means
- A61B5/742—Details of notification to user or communication with user or patient ; user input means using visual displays
- A61B5/743—Displaying an image simultaneously with additional graphical information, e.g. symbols, charts, function plots
Definitions
- the present invention relates to a program, an output device, and a data processing method for analyzing an electrocardiogram.
- an electrocardiogram sample with a flag for performing an emergency detailed examination is transmitted to a cardiologist for detailed examination.
- this ECG sample may differ from the ECG example typical of heart disease. For this reason, doctors who are not cardiologists may not be able to use the electrocardiogram sample as a basis for determining heart disease.
- the present invention has been made in view of these points.
- the object is to provide a data processing method.
- the program according to the first aspect of the present invention is stored in a computer as one or more fibrillation wave electrocardiograms containing fibrillation waves, which are included in the whole electrocardiogram data obtained by measuring the time change of the action potential accompanying the electrical activity of the heart muscle.
- a first machine learning model for classifying a plurality of electrocardiogram data containing the fibrillation waves into electrocardiogram data with the clear fibrillation waves and electrocardiogram data with the unclear fibrillation waves;
- a plurality of divided electrocardiogram data are generated by dividing the entire electrocardiogram data, and the plurality of electrocardiogram data are combined into the electrocardiogram data including the fibrillation wave and the fibrillation wave.
- input the generated plurality of divided electrocardiogram data, and output from the second machine learning model as electrocardiogram data containing fibrillation waves The output one or more fibrillation wave electrocardiogram data may be obtained.
- the clear electrocardiogram data may be output in a state in which image data indicating that fibrillation waves are included is superimposed on a time region between a plurality of R waves in the clear electrocardiogram data.
- image data indicating that a fibrillation wave is included may be superimposed on a time region within a predetermined period before the timing at which the Q wave starts in the clear electrocardiogram data.
- the image data indicating that the fibrillation wave is included in the time region from the timing when the T wave ends to the timing when the Q wave starts in the clear electrocardiogram data is superimposed. can be output.
- the program further has a step of determining the presence or absence of a signal in a frequency range corresponding to fibrillation waves by frequency-analyzing the acquired clear electrocardiogram data, and in the outputting step, The clear electrocardiogram data determined to be present in the frequency range may be output.
- the output device of the second aspect of the present invention outputs one or more fibrillation wave electrocardiogram data containing fibrillation waves, which is included in the total electrocardiogram data obtained by measuring the time change of the action potential accompanying the electrical activity of the heart muscle.
- a first acquisition unit that acquires; and a first machine learning model that classifies a plurality of electrocardiogram data containing the fibrillation waves into electrocardiogram data with clear fibrillation waves and electrocardiogram data with unclear fibrillation waves.
- a second acquisition unit that inputs the one or more fibrillation wave electrocardiogram data and acquires clear electrocardiogram data output as the electrocardiogram data with the fibrillation wave clear from the first machine learning model; an output unit that outputs the clear electrocardiogram data acquired by the second acquisition unit.
- a data processing method includes one or more fibrillation waves included in whole electrocardiogram data, which is executed by a computer and which measures changes over time in action potentials associated with electrical activity of the heart muscle.
- the data processing method comprises, before the step of acquiring the clear electrocardiogram data, obtaining a plurality of learning electrocardiogram data, which are a plurality of electrocardiogram data containing fibrillation; and the plurality of learning electrocardiograms.
- a doctor grasps an electrocardiogram waveform that serves as a basis for a doctor to determine that a patient has a heart disease.
- FIG. 1 is a diagram for explaining an outline of an electrocardiogram output system according to an embodiment; FIG. It is a figure which shows the structure of an output device.
- FIG. 3 is a diagram showing examples of whole electrocardiogram data and segmented electrocardiogram data;
- FIG. 3 is a diagram showing an example of an electrocardiogram containing fibrillation waves;
- FIG. 4 is a diagram showing an example of output of clear electrocardiogram data by an output unit;
- 4 is a flowchart showing a processing procedure for generating a machine learning model for display by an output device;
- 4 is a flowchart showing a processing procedure for outputting clear electrocardiogram data by an output device;
- FIG. 1 is a diagram for explaining an outline of an electrocardiogram output system S of this embodiment.
- the electrocardiogram output system S is a system for facilitating diagnosis by a doctor using an electrocardiogram of a patient U who may have atrial fibrillation.
- An electrocardiogram display system S includes an electrocardiograph 1 , a doctor terminal 2 , and an output device 3 .
- the electrocardiograph 1 is, for example, a Holter electrocardiograph worn by the patient U.
- the electrocardiograph 1 generates overall electrocardiogram data by measuring changes over time in action potentials associated with electrical activity of the patient's U heart muscle.
- the electrocardiograph 1 transmits the generated overall electrocardiogram data to the output device 3 via a network N including wireless communication lines.
- Time information indicating the measurement time of the overall electrocardiogram data is associated with the overall electrocardiogram data.
- the whole electrocardiogram data generated by the electrocardiograph 1 may be delivered to the output device 3 using a storage medium, for example, without going through the network N.
- the doctor terminal 2 is a terminal used by a doctor and includes, for example, a display and a computer.
- the doctor terminal 2 outputs a waveform image based on a part of the electrocardiogram data received from the output device 3 among the entire electrocardiogram data generated by the electrocardiograph 1 to the display.
- the output device 3 is, for example, a server.
- the output device 3 outputs information that assists a doctor in diagnosing whether or not the heart of the patient U is suffering from atrial fibrillation.
- the output device 3 receives the whole electrocardiogram data of the patient U from the electrocardiograph 1 or the doctor's terminal 2 .
- the output device 3 generates a plurality of divided electrocardiogram data by dividing the received whole electrocardiogram data.
- the output device 3 acquires a plurality of fibrillation wave electrocardiogram data including a fibrillation wave peculiar to atrial fibrillation among the plurality of divided electrocardiogram data.
- the output device 3 is a trained display machine learning model for classifying a plurality of electrocardiogram data including fibrillation waves into electrocardiogram data with clear fibrillation waves and electrocardiogram data with unclear fibrillation waves (see 1 machine learning model) is read from the storage unit.
- the internal configuration of the machine learning model is arbitrary, it is configured by, for example, a CNN (Convolutional Neural Network).
- the output device 3 inputs the obtained plurality of fibrillation wave electrocardiogram data to the learned machine learning model for display, and acquires clear electrocardiogram data output by the machine learning model for display as electrocardiogram data with clear fibrillation waves. .
- the output device 3 outputs the acquired clear electrocardiogram data to the doctor terminal 2 . In this way, the output device 3 enables the doctor to grasp the waveform of the electrocardiogram showing the grounds for the doctor to judge that the heart of the patient U is suffering from atrial fibrillation.
- the output device 3 is not limited to a computer separate from the doctor terminal 2 .
- the output device 3 may be the same computer as the doctor terminal 2 . The configuration and operation of the output device 3 will be described in detail below.
- FIG. 2 is a diagram showing the configuration of the output device 3.
- the output device 3 includes a communication unit 31 , a machine learning unit for judgment 32 , a machine learning unit for display 33 , a storage unit 34 and a control unit 35 .
- the control unit 35 has a first acquisition unit 351 , a second acquisition unit 352 , a determination unit 353 , an output unit 354 , a reception unit 355 and a generation unit 356 .
- the communication unit 31 has a communication controller for transmitting and receiving data between the electrocardiograph 1 and the doctor terminal 2 via the network N.
- the communication unit 31 notifies the control unit 35 of data received via the network N.
- the determination machine learning unit 32 learns based on the electrocardiogram data for learning used as teacher data, thereby dividing the input plural electrocardiogram data into electrocardiogram data including fibrillation waves and electrocardiogram data not including fibrillation waves. It functions as a judgment machine learning model (equivalent to a second machine learning model) that can be classified into electrocardiogram data.
- the decision machine learning unit 32 includes, for example, a processor that executes various calculations using CNN, and a memory that stores CNN coefficients.
- the determination machine learning unit 32 classifies the input electrocardiogram data into electrocardiogram data containing fibrillation waves and electrocardiogram data not containing fibrillation waves, and outputs the data respectively.
- the machine learning unit for display 33 learns based on the electrocardiogram data for learning used as teacher data, and converts a plurality of input electrocardiogram data including fibrillation waves into electrocardiogram data with clear fibrillation waves and fibrillation data. It functions as a machine learning model for display, as described above, that can classify waves as ambiguous ECG data.
- the display machine learning unit 33 includes, for example, a processor that executes various calculations using CNN, and a memory that stores CNN coefficients.
- the display machine learning unit 33 classifies the input electrocardiogram data including the fibrillation waves into electrocardiogram data with clear fibrillation waves and electrocardiogram data with unclear fibrillation waves, and outputs the data.
- the storage unit 34 includes storage media such as ROM (Read Only Memory), RAM (Random Access Memory), and hard disk.
- the storage unit 34 stores programs executed by the control unit 35 .
- the storage unit 34 also stores various data necessary when the control unit 35 executes various calculations.
- the control unit 35 is, for example, a CPU (Central Processing Unit).
- the control unit 35 functions as a first acquisition unit 351 , a second acquisition unit 352 , a determination unit 353 , an output unit 354 , a reception unit 355 and a generation unit 356 by executing programs stored in the storage unit 34 .
- a CPU Central Processing Unit
- the first acquisition unit 351 communicates with the electrocardiograph 1 via the communication unit 31 .
- the first acquisition unit 351 acquires, from the electrocardiograph 1, overall electrocardiogram data obtained by measuring changes over time in action potentials associated with electrical activity of the heart muscle.
- the first acquisition unit 351 acquires one or more fibrillation wave electrocardiogram data including fibrillation waves from the electrocardiogram data included in the acquired overall electrocardiogram data.
- the first acquisition unit 351 divides the acquired whole electrocardiogram data into a plurality of divided electrocardiogram data. For example, the first acquisition unit 351 generates a plurality of divided electrocardiogram data by dividing the whole electrocardiogram data.
- FIG. 3 is a diagram showing examples of whole electrocardiogram data and segmented electrocardiogram data.
- the upper side shows the whole electrocardiogram included in the whole electrocardiogram data, and the lower side shows the divided electrocardiograms contained in the divided electrocardiogram data.
- the whole electrocardiogram shows the measurement result of measuring the time change of the action potential of the heart muscle of the patient U over a predetermined measurement time. The measurement time is 24 hours as an example.
- the divided electrocardiogram data are obtained by dividing the whole electrocardiogram data. For example, the whole electrocardiogram data is divided for each rising position at which the potential indicated by the whole electrocardiogram data changes from below the threshold to above the threshold.
- the threshold value is, for example, a value that is assumed to be lower than the value assumed for the peaks other than the R wave and higher than the value assumed for the peaks of the waves other than the R wave.
- the divided electrocardiogram data may be obtained by dividing the whole electrocardiogram data into predetermined time lengths.
- the predetermined length of time is, for example, a period of time corresponding to one cycle or several cycles of the R wave included in the electrocardiogram data.
- the rightmost and leftmost Rs in the segmented electrocardiogram of FIG. 3 each indicate a part of the R wave.
- the first acquisition unit 351 acquires one or more pieces of fibrillation-wave electrocardiogram data including fibrillation waves from among the plurality of divided electrocardiogram data.
- FIGS. 4(a) to 4(c) are diagrams showing examples of electrocardiograms including fibrillation waves.
- FIG. 4(a) shows a normal electrocardiogram.
- FIG. 4(b) shows an example of an electrocardiogram containing a fibrillation wave in which the T wave is clear.
- FIG. 4(c) shows an example in which the T wave is not clear in an electrocardiogram containing fibrillation waves.
- the vertical axis of FIG. 4(a) indicates potential, and the horizontal axis of FIG. 4(a) indicates time. P, Q, R, S, and T in FIG.
- FIG. 4(a) indicate P wave, Q wave, R wave, S wave, and T wave, respectively.
- the P wave, Q wave, R wave, S wave and T wave are repeated at regular intervals.
- a normal electrocardiogram shown in FIG. 4(a) does not include fibrillation waves.
- FIG. 4(b) indicates a fibrillation wave.
- the electrocardiogram since the electrocardiogram includes fibrillation waves, it can be seen that atrial fibrillation has occurred in the patient's U heart.
- FIG. 4(b) in the electrocardiogram in the state of atrial fibrillation, the disappearance of the P wave is observed and the period of the R wave, etc. becomes irregular.
- FIG. 4(b) shows an example of an electrocardiogram containing fibrillation waves.
- the example of FIG. 4(c) differs from the example of FIG. 4(b) in that the T wave is not clear.
- the P-wave disappears, and the period of the R-wave becomes irregular.
- the first acquisition unit 351 inputs a plurality of divided electrocardiogram data to the determination machine learning unit 32 that functions as a determination machine learning model (equivalent to a second machine learning model), and the determination machine One or more pieces of fibrillation-wave electrocardiogram data output from the learning unit 32 as electrocardiogram data containing fibrillation waves are acquired.
- the first acquisition unit 351 outputs the acquired one or more fibrillation wave electrocardiogram data to the second acquisition unit 352 .
- the second acquisition unit 352 acquires clear electrocardiogram data with clear fibrillation waves from the one or more fibrillation wave electrocardiogram data acquired by the first acquisition unit 351 .
- the second acquisition unit 352 inputs one or more fibrillation wave electrocardiogram data to the display machine learning unit 33 that functions as a display machine learning model, and the display machine learning unit 33 Clear electrocardiogram data output as electrocardiogram data with clear fibrillation waves is obtained.
- the second acquisition unit 352 removes waves different from the fibrillation-wave electrocardiogram data from the fibrillation-wave electrocardiogram data. may be input to the machine learning unit 33 for display. For example, the second acquisition unit 352 removes one or more of the P wave, Q wave, R wave, S wave, and T wave (see FIGS. 4(a) to 4(c)) to obtain a modified fibrillation wave. Generate ECG data. The second acquisition unit 352 inputs the generated one or more modified fibrillation electrocardiogram data to the display machine learning unit 33 .
- the second acquisition unit 352 generates modified fibrillation electrocardiogram data by removing waves having potentials equal to or higher than a reference value from an electrocardiogram waveform containing fibrillation waves.
- the reference value is, for example, higher than the value assumed as the potential of the peak of the fibrillation wave.
- the second acquisition unit 352 identifies a time region in which the Q wave, R wave, S wave, T wave, etc. occurs, and sets the potential of the electrocardiogram to zero in this time region to obtain waves other than the fibrillation wave. may be generated to remove the modified fibrillation ECG data.
- the second acquisition unit 352 identifies a time region in which the potential is equal to or higher than the threshold as a time region in which the R wave is generated.
- the time regions in which the Q, S and T waves occurred can be deduced from the time regions in which the preceding and following R waves occurred.
- the second acquisition unit 352 identifies the time regions in which the Q wave, the S wave, and the T wave are generated based on the identified time regions in which the R waves are generated.
- the second acquisition unit 352 may generate modified fibrillation electrocardiogram data by setting the potential of the electrocardiogram to zero in the time regions where the specified Q wave, R wave, S wave, T wave, etc. occur. .
- the second acquisition unit 352 inputs one or more modified fibrillation electrocardiogram data from which waves different from the fibrillation wave are removed to the machine learning unit for display 33, and the machine learning unit for display 33 detects that the fibrillation wave is clear.
- Acquire modified fibrillation electrocardiogram data output as normal electrocardiogram data.
- the second acquisition unit 352 obtains the fibrillation-wave before removing waves different from the fibrillation-wave in the modified fibrillation-wave electrocardiogram data.
- the dynamic electrocardiogram data may be acquired as clear electrocardiogram data with a clear fibrillation wave.
- the determination unit 353 determines whether or not there is a signal in the frequency range corresponding to the fibrillation wave by frequency-analyzing the clear electrocardiogram data acquired by the second acquisition unit 352 .
- Frequency analysis is for example Fourier analysis or wavelet analysis.
- the determination unit 353 notifies the output unit 354 of the determination result of the presence/absence of the signal in the frequency range corresponding to the fibrillation wave.
- the output unit 354 communicates with the doctor terminal 2 via the communication unit 31 .
- the output unit 354 outputs the clear electrocardiogram data acquired by the second acquisition unit 352 .
- the output unit 354 outputs clear electrocardiogram data to the display of the doctor terminal 2 .
- the output unit 354 outputs clear electrocardiogram data in a state in which image data indicating that fibrillation waves are included is superimposed on the time region between the plurality of R waves in the clear electrocardiogram data obtained by the second obtaining unit 352. .
- FIG. 5 is a diagram showing an example of clear electrocardiogram data output by the output unit 354.
- FIG. 5 The image shown in FIG. 5 is output to the display D of the doctor terminal 2.
- R indicates an R wave.
- the output unit 354 displays image data M indicating that fibrillation waves are included in the time domain between multiple R waves.
- the image data M is represented by a thick elliptical frame.
- the output unit 354 displays the image data M with different magnitudes in the time direction according to the cycle of the R wave. For example, when the period of the R wave is smaller than a predetermined value, the output unit 354 displays the image data M having the first size in the time direction. The output unit 354 displays the image data M having the second size in the time direction when the period of the R wave is equal to or greater than a predetermined value. The second size is larger than the first size. In the example of FIG. 5, the output unit 354 outputs the image data M together with the message "f wave exists" indicating that the image data M corresponds to the position of the fibrillation wave.
- the output unit 354 outputs the image data M indicating that the fibrillation waves are included in the time domain between the plurality of R waves, superimposed thereon.
- the output unit 354 outputs the image data M superimposed on a time domain including intermediate positions of a plurality of R waves. In this way, the output unit 354 can make it easier for the doctor to grasp the time region in which clear fibrillation waves are included.
- the output unit 354 identifies the time region in which the P wave should occur if atrial fibrillation does not occur, and displays image data M indicating that the identified time region includes a clear fibrillation wave. good. Since the P wave occurs before the Q wave, the output unit 354 may output the image data M superimposed on the time domain within a predetermined period before the timing at which the Q wave starts. The predetermined time is determined, for example, so that the image data M includes the area of the P wave if the P wave does not disappear. In normal electrocardiogram data in which atrial fibrillation has not occurred, the P wave occurs between the timing when the T wave ends and the timing when the Q wave starts (see FIG. 4(a)). For this reason, the output unit 354 specifies the positions of the T wave and the Q wave in the electrocardiogram data, and outputs the image data M in a superimposed state from the timing when the T wave ends to the timing when the Q wave starts. good too.
- the output unit 354 may output clear electrocardiogram data, among the clear electrocardiogram data acquired by the second acquisition unit 352, which is determined by the determination unit 353 to include a signal in the frequency range corresponding to the fibrillation wave.
- the output unit 354 may not output the clear electrocardiogram data that the determination unit 353 determines that there is no signal in the frequency range corresponding to fibrillation among the clear electrocardiogram data acquired by the second acquisition unit 352 . In this way, the output unit 354 can prevent the clear electrocardiogram data from being output when the second acquisition unit 352 erroneously acquires clear electrocardiogram data that does not include fibrillation.
- the first acquisition unit 351 acquires a plurality of electrocardiogram data for learning, which are a plurality of electrocardiogram data containing fibrillation waves. At this time, the first acquisition unit 351 acquires a plurality of whole electrocardiogram data from the electrocardiographs 1 worn by the plurality of users U, respectively. The first acquisition unit 351 uses the determination machine learning unit 32 that classifies a plurality of electrocardiogram data into electrocardiogram data containing fibrillation waves and electrocardiogram data not containing fibrillation waves.
- the first acquisition unit 351 inputs the whole electrocardiogram data to the machine learning model for judgment, and converts the plurality of electrocardiogram data including the fibrillation output by the machine learning model for judgment into the plurality of electrocardiograms for learning. Get it as data.
- the reception unit 355 communicates with the doctor terminal 2 via the communication unit 31.
- the receiving unit 355 receives an instruction to classify a plurality of electrocardiogram data for learning into clear electrocardiogram data for learning with clear fibrillation waves and unclear electrocardiogram data for learning with unclear fibrillation waves. More specifically, the reception unit 355 sequentially outputs the plurality of learning electrocardiogram data acquired by the first acquisition unit 351 to the display of the doctor terminal 2 .
- the receiving unit 355 is instructed by a doctor to classify the electrocardiogram data for learning being output into either clear electrocardiogram data for learning with clear fibrillation waves or unclear electrocardiogram data for learning with unclear fibrillation waves. is received from the doctor terminal 2 for each electrocardiogram data for learning.
- the reception unit 355 classifies each of the plurality of electrocardiogram data for learning by a doctor into clear electrocardiogram data for learning with clear fibrillation waves or ambiguous electrocardiogram data for learning with unclear fibrillation waves. The result is labeled and stored in the storage unit 34.
- the determination unit 353 may determine the presence or absence of a signal in the frequency range corresponding to the fibrillation wave by frequency-analyzing the plurality of electrocardiogram data for learning acquired by the first acquisition unit 351 .
- the determination unit 353 may delete from the storage unit 34 the electrocardiogram data for learning determined as not containing the signal in the frequency range corresponding to the fibrillation wave. In this way, the determination unit 353 can prevent the learning electrocardiogram data determined not to include signals in the frequency range corresponding to fibrillation waves from being used for machine learning by the generation unit 356 .
- the generating unit 356 generates a display machine learning model that classifies a plurality of electrocardiogram data containing fibrillation waves into electrocardiogram data with clear fibrillation waves and electrocardiogram data with unclear fibrillation waves.
- the generation unit 356 performs machine learning using a plurality of learning electrocardiogram data labeled as learning clear electrocardiogram data and a plurality of learning electrocardiogram data labeled as learning unclear electrocardiogram data as teacher data. Generate a machine learning model for display.
- the generation unit 356 can generate the machine learning model for display described above.
- the electrocardiogram display system S When generating the machine learning model for display, the electrocardiogram display system S displays a plurality of fibrillation waves including fibrillation waves output by the machine learning model for judgment to the doctor who confirms whether the fibrillation waves are clear. Since the ECG data is transmitted, the number of ECG data to be confirmed by the doctor can be narrowed down. Therefore, the electrocardiogram display system S can create a machine learning model for display in a short period of time.
- FIG. 6 is a flowchart showing a processing procedure for generating a machine learning model for display by the output device 3. As shown in FIG. This processing procedure starts, for example, when the reception unit 355 receives an instruction to generate a machine learning model for display from the doctor terminal 2 .
- the first acquisition unit 351 acquires a plurality of electrocardiogram data for learning, which are a plurality of electrocardiogram data containing fibrillation (S101).
- the receiving unit 355 receives an instruction from the doctor terminal 2 to classify the plurality of electrocardiogram data for learning into clear electrocardiogram data for learning with clear fibrillation waves and unclear electrocardiogram data for learning with unclear fibrillation waves. (S102).
- the receiving unit 355 allows the doctor to classify each of the plurality of electrocardiogram data for learning into clear electrocardiogram data for learning with clear fibrillation waves or ambiguous electrocardiogram data for learning with unclear fibrillation waves.
- the classified results are labeled and stored in the storage unit 34 .
- the generation unit 356 performs machine learning using a plurality of learning electrocardiogram data labeled as learning clear electrocardiogram data and a plurality of learning electrocardiogram data labeled as learning unclear electrocardiogram data as teacher data. , a machine learning model for display is generated (S103), and the process ends.
- FIG. 7 is a flowchart showing a processing procedure for outputting clear electrocardiogram data by the output device 3. As shown in FIG. This processing procedure starts when the first acquisition unit 351 acquires from the electrocardiograph 1 the overall electrocardiogram data obtained by measuring the time change of the action potential associated with the electrical activity of the heart muscle.
- the first acquisition unit 351 acquires one or more pieces of fibrillation electrocardiogram data containing fibrillation waves, included in the acquired overall electrocardiogram data (S201).
- the second acquisition unit 352 inputs one or more fibrillation wave electrocardiogram data to the display machine learning unit 33 that functions as a display machine learning model (S202), and the display machine learning unit 33 receives fibrillation waves.
- the clear electrocardiogram data output as clear electrocardiogram data is acquired (S203).
- the output unit 354 outputs the clear electrocardiogram data acquired by the second acquisition unit 352 to the doctor terminal 2 (S204), and ends the process.
- the second acquisition unit 352 inputs the acquired plurality of fibrillation electrocardiogram data to the learned machine learning model for display, and uses the clear electrocardiogram data output by the machine learning model for display as electrocardiogram data with clear fibrillation waves. get.
- the output unit 354 outputs the identified clear electrocardiogram data to the doctor terminal 2 . In this way, the output unit 354 can allow the doctor to grasp the waveform of the electrocardiogram, which is the basis for the doctor to determine that the patient's heart is in atrial fibrillation.
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Cardiology (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- General Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
Description
図1は、本実施形態の心電図出力システムSの概要を説明するための図である。心電図出力システムSは、心房細動が生じている可能性がある患者Uの心電図を用いて医師が患者を診断しやすくするためのシステムである。心電図表示システムSは、心電計1と、医師端末2と、出力装置3とを備える。
図2は、出力装置3の構成を示す図である。出力装置3は、通信部31と、判定用機械学習部32と、表示用機械学習部33と、記憶部34と、制御部35とを有する。制御部35は、第1取得部351と、第2取得部352と、判定部353と、出力部354と、受付部355と、生成部356と、を有する。
第2取得部352は、第1取得部351が取得した一以上の細動波心電図データのうち、細動波が明瞭な明瞭心電図データを取得する。本明細書の例では、第2取得部352は、表示用機械学習モデルとして機能する表示用機械学習部33に対し、一以上の細動波心電図データを入力し、表示用機械学習部33から細動波が明瞭な心電図データとして出力された明瞭心電図データを取得する。
出力部354は、通信部31を介して、医師端末2と通信する。出力部354は、第2取得部352が取得した明瞭心電図データを出力する。例えば、出力部354は、医師端末2のディスプレイに明瞭心電図データを出力する。出力部354は、第2取得部352が取得した明瞭心電図データにおける複数のR波の間の時間領域に、細動波が含まれることを示す画像データを重ねた状態で明瞭心電図データを出力する。
図2の説明に戻る。以下、第2取得部352が明瞭心電図データを取得する前の処理として、表示用機械学習モデルを生成するための機械学習時の処理について説明する。第1取得部351は、細動波が含まれている複数の心電図データである複数の学習用心電図データを取得する。このとき、第1取得部351は、複数のユーザUにそれぞれ装着された心電計1から複数の全体心電図データを取得する。第1取得部351は、複数の心電図データを細動波が含まれている心電図データと細動波が含まれていない心電図データとに分類する判定用機械学習部32を利用することにより、取得した複数の全体心電図データのうち、細動波が含まれている複数の学習用心電図データを取得するものとする。具体的には、第1取得部351は、全体心電図データを判定用機械学習モデルに入力し、判定用機械学習モデルが出力した細動波が含まれる複数の心電図データを、複数の学習用心電図データとして取得する。
図6は、出力装置3による表示用機械学習モデルの生成の処理手順を示すフローチャートである。この処理手順は、例えば、受付部355が表示用機械学習モデルの生成の指示を医師端末2から受け付けたときに開始する。
図7は、出力装置3による明瞭心電図データの出力の処理手順を示すフローチャートである。この処理手順は、心臓筋肉の電気的活動にともなう活動電位の時間変化を測定した全体心電図データを心電計1から第1取得部351が取得したときに開始する。
第2取得部352は、取得した複数の細動波心電図データを学習済みの表示用機械学習モデルに入力し、細動波が明瞭な心電図データとして表示用機械学習モデルが出力した明瞭心電図データを取得する。出力部354は、特定した明瞭心電図データを医師端末2に出力する。このようにして、出力部354は、患者の心臓が心房細動を生じていると医師が判断するための根拠となる心電図の波形を医師が把握できるようにすることができる。
2 医師端末
3 出力装置
31 通信部
32 判定用機械学習部
33 表示用機械学習部
34 記憶部
35 制御部
351 第1取得部
352 第2取得部
353 判定部
354 出力部
355 受付部
356 生成部
Claims (9)
- コンピュータに、
心臓筋肉の電気的活動にともなう活動電位の時間変化を測定した全体心電図データに含まれる、細動波が含まれる一以上の細動波心電図データを取得するステップと、
前記細動波が含まれる複数の心電図データを、前記細動波が明瞭な心電図データと前記細動波が不明瞭な心電図データとに分類する第1機械学習モデルに対し、前記一以上の細動波心電図データを入力し、前記第1機械学習モデルから前記細動波が明瞭な前記心電図データとして出力された明瞭心電図データを取得するステップと、
前記明瞭心電図データを出力するステップと、
を実行させるプログラム。 - 前記一以上の細動波心電図データを取得するステップにおいて、前記全体心電図データを分割した複数の分割心電図データを生成し、複数の心電図データを、細動波が含まれる前記心電図データと細動波が含まれない前記心電図データとに分類する第2機械学習モデルに対し、生成した前記複数の分割心電図データを入力し、前記第2機械学習モデルから出力された細動波が含まれる心電図データとして出力された一以上の細動波心電図データを取得する、
請求項1に記載のプログラム。 - 前記出力するステップにおいて、前記明瞭心電図データにおける複数のR波の間の時間領域に、細動波が含まれることを示す画像データを重ねた状態で前記明瞭心電図データを出力する、
請求項1又は2に記載のプログラム。 - 前記出力するステップにおいて、前記明瞭心電図データにおけるQ波が開始するタイミングの前の所定期間内の時間領域に、細動波が含まれることを示す画像データを重ねた状態で出力する、
請求項1から3のいずれか一項に記載のプログラム。 - 前記出力するステップにおいて、前記明瞭心電図データにおけるT波が終了するタイミングからQ波が開始するタイミングまでの間の時間領域に、前記細動波が含まれることを示す前記画像データを重ねた状態で出力する、
請求項4に記載のプログラム。 - 取得した前記明瞭心電図データを周波数解析することにより、細動波に対応する周波数範囲の信号の有無を判定するステップをさらに有し、
前記出力するステップにおいて、細動波に対応する周波数範囲の信号が存在すると判定した前記明瞭心電図データを出力する、
請求項1から5のいずれか一項に記載のプログラム。 - 心臓筋肉の電気的活動にともなう活動電位の時間変化を測定した全体心電図データに含まれる、細動波が含まれる一以上の細動波心電図データを取得する第1取得部と、
前記細動波が含まれる複数の心電図データを、前記細動波が明瞭な心電図データと前記細動波が不明瞭な心電図データとに分類する第1機械学習モデルに対し、前記一以上の細動波心電図データを入力し、前記第1機械学習モデルから前記細動波が明瞭な前記心電図データとして出力された明瞭心電図データを取得する第2取得部と、
前記第2取得部が取得した前記明瞭心電図データを出力する出力部と、を備える、
出力装置。 - コンピュータが実行する、
心臓筋肉の電気的活動にともなう活動電位の時間変化を測定した全体心電図データに含まれる、細動波が含まれる一以上の細動波心電図データを取得するステップと、
前記細動波が含まれる複数の心電図データを、前記細動波が明瞭な心電図データと前記細動波が不明瞭な心電図データとに分類する第1機械学習モデルに対し、前記一以上の細動波心電図データを入力し、前記第1機械学習モデルから前記細動波が明瞭な前記心電図データとして出力された明瞭心電図データを取得するステップと、
前記明瞭心電図データを出力するステップと、
を有するデータ処理方法。 - 前記明瞭心電図データを取得するステップの前に、
細動波が含まれている複数の心電図データである複数の学習用心電図データを取得するステップと、
前記複数の学習用心電図データを、細動波が明瞭な学習用明瞭心電図データと、細動波が不明瞭な学習用不明瞭心電図データとに分類する指示を受け付けるステップと、
複数の前記学習用明瞭心電図データと、複数の前記学習用不明瞭心電図データと、を教師データとして学習することにより、前記細動波が含まれる複数の心電図データを、前記細動波が明瞭な心電図データと前記細動波が不明瞭な心電図データとに分類する機械学習モデルを生成するステップと、を備える、
請求項8に記載のデータ処理方法。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2022524975A JP7254399B1 (ja) | 2021-05-21 | 2021-12-22 | プログラム、出力装置及びデータ処理方法 |
EP21940901.8A EP4212099A4 (en) | 2021-05-21 | 2021-12-22 | PROGRAM, OUTPUT DEVICE AND DATA PROCESSING METHOD |
JP2023045830A JP2023075342A (ja) | 2021-05-21 | 2023-03-22 | プログラム、出力装置及びデータ処理方法 |
US18/300,391 US20230248318A1 (en) | 2021-05-21 | 2023-04-14 | Storage medium, output apparatus, and data processing method |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021-085848 | 2021-05-21 | ||
JP2021085848 | 2021-05-21 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US18/300,391 Continuation US20230248318A1 (en) | 2021-05-21 | 2023-04-14 | Storage medium, output apparatus, and data processing method |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022244291A1 true WO2022244291A1 (ja) | 2022-11-24 |
Family
ID=84140346
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/047643 WO2022244291A1 (ja) | 2021-05-21 | 2021-12-22 | プログラム、出力装置及びデータ処理方法 |
Country Status (4)
Country | Link |
---|---|
US (1) | US20230248318A1 (ja) |
EP (1) | EP4212099A4 (ja) |
JP (2) | JP7254399B1 (ja) |
WO (1) | WO2022244291A1 (ja) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002224069A (ja) * | 2001-02-07 | 2002-08-13 | Japan Science & Technology Corp | 体表面多誘導心電図装置およびそれを使用した解析方法 |
JP2008237882A (ja) * | 2007-01-11 | 2008-10-09 | Biosense Webster Inc | 心不整脈伝導性経路および病巣を特定する自動ペースマッピング |
JP2017042386A (ja) * | 2015-08-27 | 2017-03-02 | セイコーエプソン株式会社 | 生体情報処理システム及びプログラム |
JP2019502426A (ja) | 2015-12-21 | 2019-01-31 | メドトロニック モニタリング インコーポレイテッド | 不整脈発現の応急精査を行うためにフラグを付けるシステム及び方法 |
US20190059763A1 (en) * | 2017-08-25 | 2019-02-28 | Cambridge Heartwear Limited | Method of detecting abnormalities in ecg signals |
WO2020047750A1 (zh) * | 2018-09-04 | 2020-03-12 | 深圳先进技术研究院 | 心律失常的检测方法、装置、电子设备及计算机存储介质 |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
RU2496413C2 (ru) * | 2008-01-14 | 2013-10-27 | Конинклейке Филипс Электроникс Н.В. | Мониторинг мерцательной аритмии |
WO2019057676A1 (en) * | 2017-09-21 | 2019-03-28 | Koninklijke Philips N.V. | ATRIAL FIBRILLATION DETECTION USING SINGLE-LEAD ECG RECORDINGS |
US11538585B2 (en) * | 2018-06-08 | 2022-12-27 | Cambridge Heartwear Limited | Detecting abnormalities in ECG signals |
US11596344B2 (en) * | 2019-10-30 | 2023-03-07 | Tencent America LLC | Deep neural network on ECG Poincare plot for atrial fibrillation classification |
-
2021
- 2021-12-22 JP JP2022524975A patent/JP7254399B1/ja active Active
- 2021-12-22 EP EP21940901.8A patent/EP4212099A4/en active Pending
- 2021-12-22 WO PCT/JP2021/047643 patent/WO2022244291A1/ja active Application Filing
-
2023
- 2023-03-22 JP JP2023045830A patent/JP2023075342A/ja active Pending
- 2023-04-14 US US18/300,391 patent/US20230248318A1/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002224069A (ja) * | 2001-02-07 | 2002-08-13 | Japan Science & Technology Corp | 体表面多誘導心電図装置およびそれを使用した解析方法 |
JP2008237882A (ja) * | 2007-01-11 | 2008-10-09 | Biosense Webster Inc | 心不整脈伝導性経路および病巣を特定する自動ペースマッピング |
JP2017042386A (ja) * | 2015-08-27 | 2017-03-02 | セイコーエプソン株式会社 | 生体情報処理システム及びプログラム |
JP2019502426A (ja) | 2015-12-21 | 2019-01-31 | メドトロニック モニタリング インコーポレイテッド | 不整脈発現の応急精査を行うためにフラグを付けるシステム及び方法 |
US20190059763A1 (en) * | 2017-08-25 | 2019-02-28 | Cambridge Heartwear Limited | Method of detecting abnormalities in ecg signals |
WO2020047750A1 (zh) * | 2018-09-04 | 2020-03-12 | 深圳先进技术研究院 | 心律失常的检测方法、装置、电子设备及计算机存储介质 |
Non-Patent Citations (1)
Title |
---|
See also references of EP4212099A4 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2022244291A1 (ja) | 2022-11-24 |
JP7254399B1 (ja) | 2023-04-10 |
JP2023075342A (ja) | 2023-05-30 |
EP4212099A4 (en) | 2024-05-15 |
US20230248318A1 (en) | 2023-08-10 |
EP4212099A1 (en) | 2023-07-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10602942B2 (en) | Method of detecting abnormalities in ECG signals | |
JP4386235B2 (ja) | 心電図の順次比較のための方法及び装置 | |
US20230143594A1 (en) | Systems and methods for reduced lead electrocardiogram diagnosis using deep neural networks and rule-based systems | |
JP2020536693A (ja) | 人工知能自己学習に基づく心電図自動解析方法及び装置 | |
Yang et al. | 12-Lead ECG arrhythmia classification using cascaded convolutional neural network and expert feature | |
Demirel et al. | Energy-efficient real-time heart monitoring on edge–fog–cloud internet of medical things | |
US3841309A (en) | Method of analyzing cerebral electrical activity | |
EP2599436A2 (en) | Assessment of medical conditions | |
CN111839498A (zh) | 心电图分析装置及其心电图分析方法 | |
Kanna et al. | Automated defective ECG signal detection using MATLAB applications | |
Tseng et al. | Clustering analysis of aging diseases and chronic habits with multivariate time series electrocardiogram and medical records | |
US11672464B2 (en) | Electrocardiogram processing system for delineation and classification | |
WO2020221632A1 (en) | Electrocardiogram signal segmentation | |
Aggoune et al. | ECG data visualization: Combining the power of Grafana and InfluxDB | |
Merino-Monge et al. | Heartbeat detector from ECG and PPG signals based on wavelet transform and upper envelopes | |
WO2022244291A1 (ja) | プログラム、出力装置及びデータ処理方法 | |
Das et al. | Interpretation of EKG with image recognition and convolutional neural networks | |
Sheffield | Computer-aided electrocardiography | |
CN116504398A (zh) | 用于使用基于变换器的神经网络来进行心律失常预测的方法和系统 | |
NO147502B (no) | Fremgangsmaate ved analyse av magnetbaand-registrering av et ambulatorisk elektrokardiogram | |
JP7479106B1 (ja) | プログラム、出力装置及び出力方法 | |
WO2024106005A1 (ja) | プログラム、出力装置及び出力方法 | |
CN116135147A (zh) | 心电信号的处理方法及装置、电子设备及存储介质 | |
Huang et al. | Generalization challenges in ECG deep learning: Insights from dataset characteristics and attention mechanism | |
Jokić et al. | An efficient approach for heartbeat classification |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2022524975 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21940901 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2021940901 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 2021940901 Country of ref document: EP Effective date: 20230413 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |