Electrocardiogram monitoring and coronary heart disease diagnosis system

Abstract

The invention relates to the technical field of medical equipment, and discloses an electrocardiogram monitoring and coronary heart disease diagnosis system. The system comprises an electrocardio lead configuration module, an electrocardio signal acquisition module, a coronary heart disease feature calculation module, an abnormal heart beat detection module, a central processing unit and the like. The electrocardiogram lead configuration module sets a lead layout scheme and generates a signal acquisition rule; the electrocardiosignal acquisition module acquires original electrocardiosignals according to the signals; the coronary heart disease feature calculation module calculates coronary heart disease feature parameters; the abnormal heart beat detection module obtains an abnormal fluctuation index; and the central processing unit integrates the analysis data to generate diagnosis parameters. In addition, the system is also provided with a focus positioning module, a storage module and an early warning module which are respectively used for focus positioning, data storage and illness state early warning. The system can accurately collect and analyze electrocardiogram data, improve the accuracy of coronary heart disease diagnosis, realize focus positioning and early warning, and provide powerful support for clinical treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of medical equipment, and in particular to an electrocardiogram monitoring and coronary heart disease diagnosis system. Background Art

[0002] Coronary heart disease (CHD), a serious cardiovascular disease that threatens human health, carries a high incidence and mortality rate worldwide. According to statistics from the World Health Organization, CHD accounts for a significant proportion of the global death toll each year, and this number continues to rise. Therefore, accurate and timely diagnosis of CHD is crucial to improving patients' survival rates and quality of life.

[0003] Traditional ECG monitoring and coronary artery disease (CAD) diagnosis methods have numerous limitations. In terms of signal acquisition, previous ECG signal acquisition methods often lack flexibility. Most rely solely on fixed, standard leads for acquisition, with a single acquisition frequency, failing to fully capture ECG signal variations under varying conditions. For example, for patients with CAD whose symptoms are more subtle, signals acquired using conventional leads may fail to reveal subtle abnormalities, leading to missed diagnoses. Furthermore, fixed-frequency acquisition cannot rapidly increase the acquisition frequency to obtain more detailed data during a flare-up, hindering accurate diagnosis of the patient's condition.

[0004] From a diagnostic analysis perspective, traditional methods for calculating characteristic parameters for coronary artery disease typically rely solely on a single ECG feature or simple mathematical models, failing to comprehensively consider the multimodal characteristics of the ECG signal. This significantly reduces diagnostic accuracy and can easily lead to misdiagnosis. For example, in the case of early-stage coronary artery disease, ECG signal changes can involve changes in multiple dimensions, making accurate diagnosis difficult based solely on single-feature analysis.

[0005] When it comes to anomaly detection, traditional methods for detecting abnormal fluctuations are mostly based on fixed baseline standards, failing to fully account for individual differences in physiological characteristics. Everyone has different cardiac physiological characteristics, such as heart rate and ECG waveform morphology. Fixed baseline standards cannot accurately reflect an individual's true condition, leading to misjudgment of abnormal fluctuations. For example, due to long-term exercise, some athletes' heart function and ECG characteristics differ from those of ordinary people. Using traditional detection methods, their normal ECG fluctuations may be misjudged as abnormal.

[0006] Furthermore, existing diagnostic systems are deficient in lesion localization. They are unable to precisely determine the location of coronary artery disease lesions, making it difficult to meet the precise positioning requirements of clinical treatment. In actual treatment, accurate lesion localization is crucial for developing personalized treatment plans. Inaccurate positioning can lead to poor treatment outcomes and even delay disease progression.

[0007] Due to the existence of the above problems, the current ECG monitoring and coronary heart disease diagnosis technology cannot meet clinical needs. There is an urgent need for a more advanced, accurate and personalized ECG monitoring and coronary heart disease diagnosis system to improve the diagnosis level of coronary heart disease, gain precious time for patients' treatment, and reduce the mortality and disability rates of coronary heart disease. Summary of the Invention

[0008] The purpose of the present invention is to provide an electrocardiogram monitoring and coronary heart disease diagnosis system to solve the problems raised in the above background technology.

[0009] To achieve the above objectives, the present invention provides the following technical solution: an electrocardiogram monitoring and coronary heart disease diagnosis system, the system comprising:

[0010] An ECG lead configuration module, used to set the ECG lead layout scheme for the target monitored object and generate corresponding signal acquisition rules based on the lead layout;

[0011] An ECG signal acquisition module is used to acquire the original ECG signals of the lead channels from the ECG database in real time according to the signal acquisition rules, and use the time domain waveforms of all channels as the basic signal set of the monitoring dimension;

[0012] A coronary heart disease feature calculation module, configured to calculate the coronary heart disease feature parameters of the lead channel based on the pathological association map preset in the electrocardiogram database using a multimodal feature fusion algorithm, and use the feature parameters as a derived signal set of the monitoring dimension;

[0013] The heartbeat abnormality detection module is used to obtain abnormal fluctuation indicators of the lead channel through a dynamic baseline calibration method, and use the fluctuation indicators as an evaluation signal set of the monitoring dimension;

[0014] The central processing unit is used to send the lead layout plan to the coronary heart disease feature calculation module and the abnormal heartbeat detection module, send the signal acquisition rules to the electrocardiogram signal acquisition module, and integrate and analyze the signal data of all monitoring dimensions to generate coronary heart disease diagnostic parameters.

[0015] Preferably, the method of obtaining the abnormal fluctuation index of the lead channel by the dynamic baseline calibration method includes:

[0016] The target lead channel and its adjacent lead channels are used as dynamic analysis nodes, historical ECG data is acquired based on the individual physiological characteristics of the target subject, and the waveform fluctuation pattern of the dynamic analysis node in the historical period is determined;

[0017] On the basis of the waveform fluctuation pattern, and based on the morphological offset of each historical ECG data in the corresponding dynamic analysis node, calculating the abnormal baseline value of each node and the morphological deviation of each historical ECG data;

[0018] Based on the abnormal baseline value of each node and the morphological deviation of each historical ECG data, a bidirectional optimization model is used to synchronously correct the node abnormality index and signal deviation parameters to obtain the final abnormality index of each node. The final abnormality index of the node corresponding to the target lead channel is used as its abnormal fluctuation index.

[0019] Set the comparison analysis node to any one of all dynamic analysis nodes, and the comparison historical ECG data to any one of all historical data. At the same time, based on the waveform fluctuation pattern, take the offset amplitude of the comparison node in the comparison data as the independent offset, take the total offset amplitude of the comparison node in all historical data as the independent total offset, take the offset amplitude of all nodes in the comparison data as the global offset, and take the total offset amplitude of all nodes in all historical data as the global total offset. At this time, judge whether the ratio of the independent offset to the independent total offset is greater than the ratio of the global offset to the global total offset. If so, it is determined that the comparison node has significant abnormal fluctuations in the comparison data; otherwise, it is determined that there are no significant abnormal fluctuations.

[0020] Preferably, the signal acquisition rules include a basic lead group and an extended lead group; the basic lead group includes multiple standard lead types and a fixed acquisition frequency corresponding to each lead type; the extended lead group includes multiple auxiliary lead types and a dynamic acquisition frequency corresponding to each lead type.

[0021] Preferably, the step of acquiring the original ECG signal of the lead channel from the ECG database in real time according to the signal acquisition rule includes:

[0022] The ECG signal acquisition module extracts the preprocessed signal data of the lead channel from the ECG database, extracts the basic waveform signal from the preprocessed data based on each standard lead type in the basic lead group, and extracts the extended waveform signal from the preprocessed data based on each auxiliary lead type in the extended lead group, and aggregates all the basic waveforms and extended waveforms into the original ECG signal of the lead channel.

[0023] Preferably, when the pathological association map is a direct pathological relationship, it is determined whether the lead channel is associated with a typical pathological node in the electrocardiogram database; if so, the coronary heart disease characteristic parameter is the sum of the characteristic contribution values ​​of all pathological nodes associated with the lead channel; otherwise, the characteristic parameter is the baseline pathological value of the lead channel itself;

[0024] When the pathological association map is an indirect pathological relationship, the coronary heart disease characteristic parameter is the dynamic weighted sum of the characteristic contribution values ​​of all pathological nodes obtained by the lead channel through the multi-lead association path.

[0025] Preferably, the system further comprises a lesion localization module connected to the central processing unit;

[0026] The lesion localization module is used to set the target pathology type, obtain the spatiotemporal matching degree between the lead channel and the target pathology type through a spatiotemporal correlation analysis method, and use the matching degree as the positioning signal set of the monitoring dimension.

[0027] Preferably, the lesion localization module obtains the spatiotemporal matching degree between the lead channel and the target pathological type by a spatiotemporal correlation analysis method, including:

[0028] Acquire a historical abnormal conduction path set of the lead channel from a pathology model library, acquire a standard conduction path set of the target pathology type from the pathology model library, and use a spatiotemporal overlap ratio between the historical conduction path set and the standard conduction path set as a first matching degree between the lead channel and the target pathology type;

[0029] Dynamically comparing the ECG waveform characteristic curve of the lead channel within a preset time window with the standard characteristic curve of the target pathology type, and obtaining the waveform similarity between the two as a second matching degree between the lead channel and the target pathology type;

[0030] The first matching degree, the second matching degree, or a weighted comprehensive value of the first matching degree, the second matching degree, or the ...

[0031] Preferably, the central processing unit performs feature cascading or pattern reorganization on the signal data of all monitoring dimensions to generate the coronary heart disease diagnostic parameters.

[0032] Preferably, the system further comprises a storage module connected to the central processing unit, and the storage module is used to store the original electrocardiogram signal of the lead channel, coronary heart disease characteristic parameters, abnormal fluctuation index and coronary heart disease diagnostic parameters.

[0033] Preferably, the system also includes an early warning module connected to the central processing unit, and the early warning module is used to generate multi-level early warning instructions based on the comparison results of the diagnostic parameters and the preset risk threshold, and send the early warning instructions to the designated medical terminal through the communication interface.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] At the signal acquisition level, the ECG lead configuration module can set a variety of lead layout schemes and generate corresponding signal acquisition rules, which include basic lead groups and extended lead groups. The multiple standard lead types and fixed acquisition frequencies of the basic lead group can ensure the acquisition of conventional ECG signals; the auxiliary lead types and dynamic acquisition frequencies of the extended lead group can flexibly adjust the acquisition strategy according to actual conditions. For example, when a patient shows suspected symptoms, the dynamic acquisition frequency can quickly increase the acquisition frequency, comprehensively and accurately acquire ECG signals, avoid missing key information, and provide sufficient and high-quality data support for subsequent diagnosis.

[0036] The CHD feature calculation module calculates CHD characteristic parameters based on a pre-set pathology association map using a multimodal feature fusion algorithm. When the pathology association map indicates a direct pathology relationship, it accurately determines the association between lead channels and pathology nodes and rationally calculates characteristic parameters. When the pathology relationship is indirect, it dynamically weights the sum of characteristic contribution values ​​across multiple lead association paths, fully accounting for the complex connections between multimodal features of ECG signals. This significantly improves the accuracy of CHD characteristic parameter calculation, thereby enhancing the reliability of diagnostic results and reducing the incidence of misdiagnosis.

[0037] The heartbeat anomaly detection module utilizes a dynamic baseline calibration method, using the target lead channel and its adjacent lead channels as dynamic analysis nodes. It acquires historical ECG data based on individual physiological characteristics and determines waveform fluctuation patterns. Based on this, it calculates abnormal baseline values, morphological deviation, and other parameters. A bidirectional optimization model is then used to simultaneously correct node anomaly indicators and signal deviation parameters. This approach fully accounts for individual differences, more accurately capturing abnormal fluctuation indicators and effectively preventing normal physiological fluctuations from being misclassified as abnormal, thereby improving the accuracy of anomaly detection.

[0038] The lesion localization module uses spatiotemporal correlation analysis to integrate the spatiotemporal overlap ratio between the historical abnormal conduction pathway set and the standard conduction pathway set, as well as the waveform similarity between the ECG waveform characteristic curve and the standard characteristic curve, to determine the spatiotemporal matching degree between the lead channel and the target pathology type. This function can accurately determine the location of coronary artery disease lesions, providing a key basis for doctors to develop personalized treatment plans, significantly improving the targetedness and effectiveness of treatment, and enhancing patient outcomes.

[0039] The central processing unit integrates and analyzes signal data from various monitoring dimensions. Whether through feature concatenation or pattern recombining to generate CHD diagnostic parameters, it fully exploits the potential information within the ECG data. The storage module stores raw ECG signals, CHD characteristic parameters, and other data for subsequent review and comparative analysis. The early warning module compares diagnostic parameters with pre-set risk thresholds, generates multi-level warning instructions, and sends them to designated medical terminals. This allows medical staff to promptly monitor changes in the patient's condition so that appropriate measures can be taken to protect the patient's health and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a working principle diagram of the electrocardiogram monitoring and coronary heart disease diagnosis system of the present invention;

[0041] Figure 2 A diagram showing the working principle of obtaining abnormal fluctuation indicators for the dynamic baseline calibration method;

[0042] Figure 3 The working principle diagram for signal acquisition rules and original ECG signal acquisition;

[0043] Figure 4 The working principle diagram for calculating characteristic parameters of coronary heart disease;

[0044] Figure 5 This is the working principle diagram of the lesion localization module. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] See also Figure 1-Figure 5 The present invention provides an electrocardiogram monitoring and coronary heart disease diagnosis system, and the specific implementation methods are as follows:

[0047] The system comprises an ECG lead configuration module, an ECG signal acquisition module, a coronary artery disease signature calculation module, an abnormal heartbeat detection module, and a central processing unit. The ECG lead configuration module determines the appropriate ECG lead layout based on the individual's specific circumstances, such as age, gender, and physical condition. Once the layout is determined, corresponding signal acquisition rules are generated based on the plan. These rules specify which lead channels to collect signals from, the frequency of acquisition, and the interval between acquisitions.

[0048] The ECG signal acquisition module, following the signal acquisition rules generated by the ECG lead configuration module, acquires raw ECG signals from the ECG database in real time. After acquiring the time-domain waveforms for all channels, they serve as the basic signal set for monitoring. This basic signal set contains essential information about the target subject's ECG activity, providing raw data support for subsequent analysis.

[0049] The CHD feature calculation module uses a multimodal feature fusion algorithm to calculate CHD characteristic parameters for each lead channel based on pre-defined pathology association maps in the ECG database. By analyzing various features related to CHD in the ECG signal, the calculated characteristic parameters are used as a derived signal set for the monitoring dimension. These parameters can reflect CHD-related characteristic information in the ECG signal from different perspectives.

[0050] The heartbeat anomaly detection module uses a dynamic baseline calibration method to obtain abnormal fluctuation indicators for lead channels, which effectively reflect abnormal changes in ECG signals. These abnormal fluctuation indicators serve as the evaluation signal set for monitoring dimensions, providing an important basis for determining whether there are any heart anomalies.

[0051] The central processing unit (CPU) plays a core coordination and analysis role. It sends the lead layout scheme generated by the ECG lead configuration module to the coronary artery disease feature calculation module and the heartbeat abnormality detection module, enabling these two modules to perform relevant calculations and detection based on the correct lead layout. Simultaneously, it sends the signal acquisition rules to the ECG signal acquisition module to ensure that the acquisition module accurately captures the original ECG signal. Finally, the CPU integrates and analyzes the signal data from all monitoring dimensions—the basic signal set, the derived signal set, and the evaluation signal set—taking into account various factors to generate CHD diagnostic parameters, providing a key basis for the diagnosis of CHD.

[0052] The present invention will be further described below in conjunction with Examples 1 to 5:

[0053] Example 1:

[0054] During actual system operation, the dynamic baseline calibration method obtains abnormal fluctuation indicators for lead channels as follows: The target lead channel and its adjacent lead channels are selected as dynamic analysis nodes. Individual physiological characteristics of the target subject, such as height, weight, and blood pressure, are collected. This information is then used to obtain historical ECG data. This historical ECG data is analyzed to determine the waveform fluctuation pattern of the dynamic analysis nodes over the historical period. This pattern reflects the fluctuation pattern of these nodes during normal ECG activity for the subject.

[0055] Based on the morphological deviation of each historical ECG data point in the corresponding dynamic analysis node, the abnormal baseline value and the morphological deviation degree of each historical ECG data point are calculated for each node. The abnormal baseline value serves as a reference for determining whether subsequent ECG data is abnormal, while the morphological deviation degree quantifies the degree to which each historical ECG data point differs from the normal fluctuation pattern.

[0056] Next, based on each node's abnormal baseline value and the morphological deviation of each historical ECG data set, a bidirectional optimization model is used to simultaneously correct the node abnormality index and signal deviation parameters. After multiple iterations of optimization, the final abnormality index for each node is obtained, and the final abnormality index of the node corresponding to the target lead channel is used as its abnormal fluctuation index.

[0057] When judging whether there is a significant abnormal fluctuation, the comparison analysis node is set to any one of all dynamic analysis nodes, and the comparison historical ECG data is set to any one of all historical data. Based on the determined waveform fluctuation pattern, the offset amplitude of the comparison node in the comparison data (independent offset), the total offset amplitude of the comparison node in all historical data (independent total offset), the offset amplitude of all nodes in the comparison data (global offset), and the total offset amplitude of all nodes in all historical data (global total offset) are calculated respectively. Then, the ratio of the independent offset to the independent total offset and the ratio of the global offset to the global total offset are compared. If the former is greater than the latter, it is determined that the comparison node has a significant abnormal fluctuation in the comparison data, otherwise it is determined that there is no significant abnormal fluctuation. In this way, abnormal fluctuations in the ECG signal can be detected more accurately.

[0058] Consider a 65-year-old male patient with a moderate degree of cardiovascular disease who is using this system for ECG monitoring. Lead V2 in the patient's ECG monitoring system is selected as the target lead channel, with adjacent leads V1 and V3 serving as dynamic analysis nodes. Using the hospital's past examination records and historical data collected from the patient's wearable device, we obtain the patient's individual physiological characteristics, such as height, weight, and heart rate. We then extract the patient's historical ECG data, measured daily over the past week, from the ECG database.

[0059] Through in-depth analysis of these historical ECG data, waveform fluctuation curves for leads V1, V2, and V3 were plotted during historical periods (e.g., between 9:00 AM and 10:00 AM daily). By summarizing these curves' characteristics, such as shape and amplitude variations, their waveform fluctuation patterns were determined. Under this pattern, the peak amplitude of the V2 waveform during a normal cardiac cycle ranges from 0.5 to 1.0 mV, while the trough amplitude ranges from -0.1 to 0.1 mV.

[0060] The abnormal baseline value and morphological deviation are calculated based on the morphological offset of each historical ECG data in the corresponding dynamic analysis node. For example, in a certain historical ECG data, the peak amplitude of the V2 lead reached 1.3mV, which resulted in an offset of 0.3mV compared to the peak amplitude upper limit of 1.0mV in the normal fluctuation mode. Based on the offset statistical results of a large amount of historical data, a certain calculation method (no formula is involved here) is used to determine that the abnormal baseline value of the V2 lead is 0.2mV. The morphological deviation of the historical ECG data is obtained by comparing the offset with relevant data such as the abnormal baseline value (the specific calculation process does not involve formulas).

[0061] Next, a bidirectional optimization model is used to simultaneously correct node anomaly indicators and signal deviation parameters. This model comprehensively considers factors such as the morphological deviation of multiple historical ECG data and the interrelationships between nodes. After multiple iterative calculations (the specific calculation process is not included in the formula), the final anomaly indicator for the node corresponding to lead V2 is obtained.

[0062] To determine whether significant abnormal fluctuations exist, assume that lead V1 is set as the comparison analysis node and select ECG data from a specific time (e.g., yesterday's 9:30 AM) as the historical ECG data for comparison. Based on the determined waveform fluctuation pattern, the offset amplitude of lead V1 in this comparison data (independent offset) is measured, assuming it is 0.4 mV. The total offset amplitude of lead V1 across all historical data (independent total offset) is calculated, assuming it is 2.0 mV. Simultaneously, the offset amplitudes of all nodes (i.e., leads V1, V2, and V3) in this comparison data (global offset) are calculated, assuming the total is 1.0 mV. The total offset amplitude of all nodes across all historical data (global total offset) is calculated to be 5.0 mV. The ratio of the independent offset to the independent total offset is calculated to be 0.4 mV ÷ 2.0 mV = 0.2, and the ratio of the global offset to the global total offset is 1.0 mV ÷ 5.0 mV = 0.2. Since these ratios are equal, it is determined that lead V1 does not have significant abnormal fluctuations in this comparison data.

[0063] Example 2:

[0064] Signal acquisition rules play a key role in the system and include a basic lead set and an extended lead set. The basic lead set consists of multiple standard lead types, such as the common Lead I, Lead II, and Lead III, and each standard lead type corresponds to a fixed acquisition frequency. This is because standard leads have been proven in long-term clinical applications to reflect the basic characteristics of cardiac electrical activity, and a fixed acquisition frequency ensures the stability and comparability of the acquired ECG signals.

[0065] The extended lead group includes multiple auxiliary lead types, such as V1-V6, each of which corresponds to a dynamic acquisition frequency. The dynamic acquisition frequency setting is adjusted based on different monitoring needs and the specific conditions of the target patient. For example, for patients with special conditions, ECG signal changes in certain auxiliary leads may be more critical. In this case, the acquisition frequency of these auxiliary leads can be appropriately increased to observe ECG signal changes in more detail.

[0066] When the ECG signal acquisition module acquires raw ECG signals from the ECG database according to the signal acquisition rules, it first extracts the preprocessed signal data for the lead channel from the database. Then, based on each standard lead type in the basic lead group, it extracts the basic waveform signal from the preprocessed data. For the extended lead group, it similarly extracts the extended waveform signal from the preprocessed data based on each auxiliary lead type. Finally, all the extracted basic waveforms and extended waveforms are combined to obtain the raw ECG signal for the lead channel. The raw ECG signal obtained in this way contains both the basic waveform reflecting the basic electrical activity of the heart and the extended waveform that can provide more detailed information, providing comprehensive data support for subsequent analysis.

[0067] Taking the monitoring of a patient suspected of coronary artery disease as an example, the signal acquisition rules set by the system's ECG lead configuration module include a basic lead group consisting of standard leads I, II, and III, all with a fixed acquisition frequency of 1000 Hz. This is because, in clinical practice, a 1000 Hz acquisition frequency can clearly capture the ECG signal details of these standard leads, meeting the needs of preliminary analysis of the basic electrical activity characteristics of the heart. For example, lead I primarily reflects changes in the potential difference between the left and right atria of the heart. At a 1000 Hz acquisition frequency, information such as the morphology, amplitude, and time interval of the P wave, QRS complex, and T wave can be accurately recorded.

[0068] The extended lead set includes leads V1-V6 as auxiliary leads. For this patient suspected of coronary artery disease, considering the possibility of regional myocardial ischemia, the dynamic acquisition frequency of leads V1-V3 was increased to 2000Hz, and the dynamic acquisition frequency of leads V4-V6 was set to 1500Hz to more closely observe the ECG changes of the anterior and lateral walls of the heart.

[0069] When the ECG signal acquisition module begins operation, it first extracts preprocessed signal data for the patient's lead channel from the ECG database. For the basic lead group, the basic waveform signals for Leads I, II, and III are extracted from the preprocessed data at a frequency of 1000Hz. For example, from the preprocessed data for Lead I, a data point is extracted every 1ms (the time interval at a 1000Hz acquisition frequency). By connecting these data points, the basic waveform signal for Lead I is obtained, which clearly shows the changes in the heart's potential during each cardiac cycle.

[0070] For the extended lead group, the extended waveform signal is extracted according to the respective dynamic acquisition frequencies. Taking the V1 lead as an example, since its acquisition frequency is 2000Hz, a data point is extracted every 0.5ms from the preprocessed data to obtain the extended waveform signal of the V1 lead. Compared with the waveform at a lower acquisition frequency, this extended waveform signal can more accurately present subtle changes in the V1 lead ECG signal, such as abnormal conditions such as slight elevation or depression of the ST segment. Finally, all the extracted basic waveforms (leads I, II, and III) and extended waveforms (leads V1-V6) are summarized to form the original ECG signal of the patient's lead channel, providing a comprehensive and detailed data basis for subsequent diagnosis and analysis of coronary heart disease.

[0071] Example 3:

[0072] When calculating CHD characteristic parameters, the CHD feature calculation module performs different calculations based on the pathology association map. When the pathology association map indicates a direct pathology relationship, the module first determines whether the lead channel is associated with a typical pathology node in the ECG database. If the lead channel is associated with a typical pathology node, the CHD characteristic parameter is the sum of the feature contribution values ​​of all pathology nodes associated with that lead channel. This is because these associated pathology nodes are directly related to CHD, and their combined feature contribution values ​​can better reflect the degree of association between the lead channel and CHD.

[0073] If a lead channel isn't associated with a typical pathological node, the CHD characteristic parameter is the baseline pathological value for that lead channel. This baseline pathological value is derived from statistical analysis of a large amount of normal and diseased ECG data. It represents the characteristic values ​​of the lead channel under normal and commonly diseased conditions. It serves as a reference standard for assessing CHD characteristics when there's no direct association with a typical pathological node.

[0074] When the pathology association map represents an indirect pathology relationship, the CHD characteristic parameter is the dynamically weighted sum of the characteristic contribution values ​​of all pathology nodes acquired by the lead channel through the multi-lead association path. Because this is an indirect relationship, different multi-lead association paths have varying degrees of influence on the CHD characteristics, necessitating a dynamic weighted calculation of the characteristic contribution values ​​of each pathology node. Weights are determined based on factors such as the closeness of the association between different leads and the strength of signal transmission. This approach more accurately reflects the CHD characteristic parameters of the lead channel under indirect pathology relationships.

[0075] Assume that when a 58-year-old female patient is diagnosed with coronary heart disease, the coronary heart disease feature calculation module starts working. The patient's electrocardiogram (ECG) data is stored in an ECG database, and a pathology correlation map is preset in the database.

[0076] When the pathology association map is a direct pathology relationship, lead V5 is used as an example for analysis. A database query revealed that lead V5 is associated with two typical pathology nodes: the "myocardial ischemia node" and the "arrhythmia node." The feature contribution value of the "myocardial ischemia node" is 0.6, and the feature contribution value of the "arrhythmia node" is 0.3. According to the calculation rules, the coronary heart disease characteristic parameter C of this lead channel is the sum of the feature contribution values ​​of all associated pathology nodes, that is, C = 0.6 + 0.3 = 0.9, where C represents the coronary heart disease characteristic parameter of lead V5, 0.6 is the feature contribution value of the "myocardial ischemia node," and 0.3 is the feature contribution value of the "arrhythmia node."

[0077] If a lead (such as lead aVR) is not associated with a typical pathological node in the database, the CHD characteristic parameter for that lead is its own baseline pathological value. Assume that the baseline pathological value for lead aVR is 0.1. This is derived through statistical analysis of a large amount of normal and diseased ECG data and represents the CHD-related characteristic value of lead aVR under normal circumstances.

[0078] When the pathology association map is an indirect pathology relationship, let's take Lead II as an example. Lead II is associated with multiple pathology nodes through multi-lead association paths. Assume that Lead II is connected to the "Inferior Myocardial Infarction Node" and the "Atrioventricular Block Node" through association paths with Lead III and Lead aVF. Based on factors such as the closeness of the association between Lead II and these leads and the strength of signal transmission, the feature contribution weight of the "Inferior Myocardial Infarction Node" is determined to be 0.7, and the feature contribution weight of the "Atrioventricular Block Node" is determined to be 0.3. The feature contribution weight of the "Inferior Myocardial Infarction Node" is 0.5, and the feature contribution weight of the "Atrioventricular Block Node" is 0.4. The CHD characteristic parameter D for lead II is the dynamically weighted sum of the characteristic contribution values ​​of all pathological nodes obtained through the multi-lead association path: D = 0.5 × 0.7 + 0.4 × 0.3 = 0.35 + 0.12 = 0.47, where D represents the CHD characteristic parameter for lead II, 0.5 represents the characteristic contribution value of the "inferior myocardial infarction node," and 0.7 represents its weight; 0.4 represents the characteristic contribution value of the "atrioventricular block node," and 0.3 represents its weight. This dynamic weighted calculation method more accurately reflects the CHD characteristic parameters of lead channels under indirect pathological relationships, providing a more reliable basis for CHD diagnosis.

[0079] Example 4:

[0080] The system also includes a lesion localization module, which is connected to the central processing unit. In practical applications, the lesion localization module first sets the target pathology type, for example, determining whether to search for specific coronary heart disease-related pathology types such as myocardial ischemia and myocardial infarction.

[0081] The spatiotemporal matching degree between the lead channel and the target pathology type is then determined through a spatiotemporal correlation analysis method. Specifically, a set of historical abnormal conduction pathways for the lead channel is obtained from the pathology model library, along with a set of standard conduction pathways for the target pathology type. The spatiotemporal overlap ratio between the historical and standard conduction pathways is used as the first degree of matching between the lead channel and the target pathology type. The higher the spatiotemporal overlap ratio, the more similar the historical abnormal conduction patterns of the lead channel are to the standard conduction pathways of the target pathology type, indicating a higher spatial and temporal correlation between the lead channel and the target pathology type.

[0082] The ECG waveform characteristic curve of the lead channel within a preset time window is dynamically compared with the standard characteristic curve of the target pathology type. The waveform similarity between the two is obtained as the second matching degree between the lead channel and the target pathology type. The ECG waveform characteristic curve contains information such as the ECG signal's changing trend and amplitude over a period of time. By comparing it with the standard characteristic curve, the degree of difference between the lead channel's ECG waveform and the target pathology type can be intuitively seen. The higher the similarity, the greater the correlation between the two.

[0083] Finally, the first matching degree, the second matching degree, or a weighted composite value of the two is used as the spatiotemporal matching degree. Depending on the actual situation, different weights can be assigned to the first and second matching degrees. By comprehensively considering both the conduction pathway and waveform characteristics, a more accurate spatiotemporal matching degree can be obtained, providing a strong basis for localizing coronary artery disease lesions.

[0084] Consider a 70-year-old male patient using the system for ECG monitoring and coronary artery disease diagnosis. The lesion localization module kicks in. First, the doctor sets the target pathology type as "myocardial infarction" based on the patient's symptoms and preliminary examination results.

[0085] The lesion localization module retrieves a set of historical abnormal conduction pathways for the patient's lead channels from the pathology model library. For example, a query revealed that abnormal conduction pathways occurred in leads V3-V5 during certain periods of the patient's past ECG monitoring data. These pathways indicate changes in the sequence and timing of ECG signal conduction. Simultaneously, the module retrieves a set of standard conduction pathways for the target pathology type, "myocardial infarction," from the pathology model library. This set of standard conduction pathways indicates that when myocardial infarction occurs, typical abnormal pathways will occur in ECG signal conduction in specific areas of the heart.

[0086] Compare and analyze the historical abnormal conduction pathways in leads V3-V5 with the standard conduction pathways for "myocardial infarction" to calculate their spatial and temporal overlap. Assuming that a detailed comparison and calculation reveals a 30% spatial overlap and a 40% temporal overlap between the historical abnormal conduction pathways in leads V3-V5 and the standard conduction pathways, and considering both temporal and spatial factors, the temporal and spatial overlap is 35%. This ratio represents the initial degree of match between leads V3-V5 and the target pathology type of "myocardial infarction."

[0087] Next, the ECG waveform characteristic curve of leads V3-V5 within a preset time window (such as within 10 minutes of the most recent ECG monitoring) is dynamically compared with the standard characteristic curve of "myocardial infarction". The standard characteristic curve shows that when myocardial infarction occurs, the ECG waveform will show typical changes such as ST segment elevation and T wave inversion. Using a professional waveform analysis algorithm, the ECG waveform characteristic curve of leads V3-V5 is compared point by point with the standard characteristic curve, and the similarity between the two is calculated. Assuming that after calculation, the similarity between the ECG waveform of leads V3-V5 and the standard characteristic curve is 60%, this is the second matching degree between leads V3-V5 and the target pathological type of "myocardial infarction".

[0088] Finally, based on actual clinical experience and a comprehensive assessment of the patient's condition, the doctor determined that the first and second matching degrees were equally important in assessing lesion localization. Therefore, a weighted combination of the two was performed, resulting in a spatiotemporal matching degree between leads V3-V5 and the target pathology type of "myocardial infarction" of (35% + 60%) ÷ 2 = 47.5%. This spatiotemporal matching degree allowed the doctor to preliminarily determine that the patient's myocardial infarction lesion was likely highly correlated with the cardiac region corresponding to leads V3-V5, providing an important reference for further diagnosis and treatment.

[0089] Example 5:

[0090] The system also includes a storage module and an early warning module, both connected to the central processing unit. The storage module plays a crucial role in data storage during system operation, storing the raw ECG signals from the lead channels. These raw ECG signals are fundamental to the entire diagnostic process, and subsequent analysis and diagnosis rely on them. Therefore, storing raw ECG signals ensures data integrity and traceability, facilitating reanalysis when needed.

[0091] The storage module also stores characteristic parameters for coronary artery disease (CAD). These parameters, derived through complex calculations, reflect the characteristics of ECG signals related to CAD and are crucial for diagnosing CAD. Abnormal fluctuation indicators are also stored in the storage module. They visually reflect abnormal changes in ECG signals and are crucial for determining whether there are any heart abnormalities.

[0092] In addition, the storage module stores the coronary heart disease diagnostic parameters generated by the central processing unit. These diagnostic parameters are obtained by comprehensively analyzing the signal data of each monitoring dimension and are quantitative assessment results of whether the target object has coronary heart disease and the degree of the disease.

[0093] The early warning module generates multi-level early warning instructions based on the comparison results of the diagnostic parameters generated by the central processing unit with the preset risk threshold. The preset risk threshold is determined based on a large amount of clinical data and medical research, and different thresholds correspond to different risk levels. When the diagnostic parameters exceed the corresponding risk threshold, the early warning module will generate early warning instructions of the corresponding level. For example, when the diagnostic parameters reach the first-level risk threshold, a first-level early warning instruction is generated, indicating that the patient may have a mild risk of coronary heart disease; when a higher-level risk threshold is reached, a higher-level early warning instruction is generated, indicating that the patient has a higher risk of coronary heart disease. The early warning module sends these early warning instructions to designated medical terminals, such as doctors' computers, mobile phones and other devices, through the communication interface, so that doctors can understand the patient's condition in a timely manner and take appropriate treatment measures.

[0094] Taking a 45-year-old male patient who was admitted to the hospital due to chest pain as an example, during the operation of the system, the storage module and the early warning module work together.

[0095] After analyzing and processing the patient's ECG data, the ECG signal acquisition module, the coronary artery disease characteristic calculation module, the heartbeat abnormality detection module, and the lesion localization module generate a large amount of data. The storage module begins storing this data, starting with the raw ECG signals from the lead channels. For example, during 24 hours of continuous ECG monitoring, the patient's 12-lead ECG signals (I, II, III, aVR, aVL, aVF, V1-V6) were collected. These signals are stored in digital form in the storage module, encoded according to international ECG data storage standards to ensure data accuracy and compatibility.

[0096] The CHD characteristic parameters calculated by the CHD characteristic calculation module for each lead channel are also stored. For example, the CHD characteristic parameter for lead V2 is 0.5, reflecting the characteristic information related to CHD in the V2 ECG signal. The storage module stores this parameter along with information such as the lead ID and calculation time for easy subsequent query and analysis.

[0097] The abnormal fluctuation index obtained by the abnormal heartbeat detection module is also stored. For example, the abnormal fluctuation index of the aVF lead calculated by the dynamic baseline calibration method is 0.3, indicating that the ECG signal of the aVF lead has a certain degree of abnormal fluctuation. The storage module records this index and its related analysis data.

[0098] The CHD diagnostic parameters generated by the central processing unit are also stored in the storage module. Assume that after integrating and analyzing the signal data from all monitoring dimensions, the patient's CHD diagnostic parameter is 0.6. This parameter comprehensively reflects the patient's likelihood of CHD, and the storage module securely stores it.

[0099] The early warning module generates multi-level early warning instructions based on the comparison results of the diagnostic parameters generated by the central processing unit with the preset risk thresholds. Based on a large amount of clinical data and expert experience, the hospital sets the first-level risk threshold for coronary heart disease diagnostic parameters to 0.5 and the second-level risk threshold to 0.7. The patient's coronary heart disease diagnostic parameter is 0.6, which is between the first-level and second-level risk thresholds, and the early warning module generates a first-level early warning instruction. The early warning module sends this early warning instruction to the mobile terminal of the doctor responsible for the patient through the hospital's internal network communication interface. After receiving the early warning information, the doctor can promptly review the patient's electrocardiogram monitoring report and diagnostic parameters, further evaluate the patient's condition, and take appropriate treatment measures if necessary, such as arranging further examinations or adjusting the treatment plan to protect the patient's health.

[0100] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0101] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An electrocardiogram monitoring and coronary heart disease diagnosis system, characterized in that: include: An ECG lead configuration module, used to set the ECG lead layout scheme for the target monitored object and generate corresponding signal acquisition rules based on the lead layout; An ECG signal acquisition module is used to acquire the original ECG signals of the lead channels from the ECG database in real time according to the signal acquisition rules, and use the time domain waveforms of all channels as the basic signal set of the monitoring dimension; A coronary heart disease feature calculation module, configured to calculate the coronary heart disease feature parameters of the lead channel based on the pathological association map preset in the electrocardiogram database using a multimodal feature fusion algorithm, and use the feature parameters as a derived signal set of the monitoring dimension; The heartbeat abnormality detection module is used to obtain abnormal fluctuation indicators of the lead channel through a dynamic baseline calibration method, and use the fluctuation indicators as an evaluation signal set of the monitoring dimension; The central processing unit is used to send the lead layout plan to the coronary heart disease feature calculation module and the abnormal heartbeat detection module, send the signal acquisition rules to the electrocardiogram signal acquisition module, and integrate and analyze the signal data of all monitoring dimensions to generate coronary heart disease diagnostic parameters.

2. The diagnostic system according to claim 1, wherein: The abnormal fluctuation index of the lead channel obtained by the dynamic baseline calibration method includes: The target lead channel and its adjacent lead channels are used as dynamic analysis nodes, historical ECG data is acquired based on the individual physiological characteristics of the target subject, and the waveform fluctuation pattern of the dynamic analysis node in the historical period is determined; On the basis of the waveform fluctuation pattern, and based on the morphological offset of each historical ECG data in the corresponding dynamic analysis node, calculating the abnormal baseline value of each node and the morphological deviation of each historical ECG data; Based on the abnormal baseline value of each node and the morphological deviation of each historical ECG data, a bidirectional optimization model is used to synchronously correct the node abnormality index and signal deviation parameters to obtain the final abnormality index of each node. The final abnormality index of the node corresponding to the target lead channel is used as its abnormal fluctuation index. Set the comparison analysis node to any one of all dynamic analysis nodes, and the comparison historical ECG data to any one of all historical data. At the same time, based on the waveform fluctuation pattern, take the offset amplitude of the comparison node in the comparison data as the independent offset, take the total offset amplitude of the comparison node in all historical data as the independent total offset, take the offset amplitude of all nodes in the comparison data as the global offset, and take the total offset amplitude of all nodes in all historical data as the global total offset. At this time, judge whether the ratio of the independent offset to the independent total offset is greater than the ratio of the global offset to the global total offset. If so, it is determined that the comparison node has significant abnormal fluctuations in the comparison data; otherwise, it is determined that there are no significant abnormal fluctuations.

3. The diagnostic system according to claim 1, wherein: The signal acquisition rules include a basic lead group and an extended lead group; the basic lead group includes multiple standard lead types and a fixed acquisition frequency corresponding to each lead type; the extended lead group includes multiple auxiliary lead types and a dynamic acquisition frequency corresponding to each lead type.

4. The diagnostic system according to claim 2, wherein: The step of acquiring the original ECG signal of the lead channel from the ECG database in real time according to the signal acquisition rule includes: The ECG signal acquisition module extracts the preprocessed signal data of the lead channel from the ECG database, extracts the basic waveform signal from the preprocessed data based on each standard lead type in the basic lead group, and extracts the extended waveform signal from the preprocessed data based on each auxiliary lead type in the extended lead group, and aggregates all the basic waveforms and extended waveforms into the original ECG signal of the lead channel.

5. The diagnostic system according to claim 1, wherein: When the pathological association map is a direct pathological relationship, determine whether the lead channel is associated with a typical pathological node in the ECG database. If so, the coronary heart disease characteristic parameter is the sum of the characteristic contribution values ​​of all pathological nodes associated with the lead channel; otherwise, the characteristic parameter is the baseline pathological value of the lead channel itself; When the pathological association map is an indirect pathological relationship, the coronary heart disease characteristic parameter is the dynamic weighted sum of the characteristic contribution values ​​of all pathological nodes obtained by the lead channel through the multi-lead association path.

6. The diagnostic system according to claim 1, wherein: Also included is a lesion localization module connected to the central processing unit; The lesion localization module is used to set the target pathology type, obtain the spatiotemporal matching degree between the lead channel and the target pathology type through a spatiotemporal correlation analysis method, and use the matching degree as the positioning signal set of the monitoring dimension.

7. The diagnostic system according to claim 6, characterized in that The lesion localization module obtains the spatiotemporal matching degree between the lead channel and the target pathological type through a spatiotemporal correlation analysis method, including: Acquire a historical abnormal conduction path set of the lead channel from a pathology model library, acquire a standard conduction path set of the target pathology type from the pathology model library, and use a spatiotemporal overlap ratio between the historical conduction path set and the standard conduction path set as a first matching degree between the lead channel and the target pathology type; Dynamically comparing the ECG waveform characteristic curve of the lead channel within a preset time window with the standard characteristic curve of the target pathology type, and obtaining the waveform similarity between the two as a second matching degree between the lead channel and the target pathology type; The first matching degree, the second matching degree, or a weighted comprehensive value of the first matching degree, the second matching degree, or the ...

8. The diagnostic system according to claim 1, wherein: The central processing unit performs feature cascading or pattern reorganization on the signal data of all monitoring dimensions to generate the coronary heart disease diagnostic parameters.

9. The diagnostic system according to claim 1, wherein: It also includes a storage module connected to the central processing unit, and the storage module is used to store the original electrocardiogram signal of the lead channel, coronary heart disease characteristic parameters, abnormal fluctuation indicators and coronary heart disease diagnostic parameters.

10. The diagnostic system according to claim 1, wherein: It also includes an early warning module connected to the central processing unit, which is used to generate multi-level early warning instructions based on the comparison results of the diagnostic parameters and the preset risk threshold, and send the early warning instructions to the designated medical terminal through the communication interface.

Images