Electrocardiogram-based left ventricular dysfunction and ejection fraction monitoring

Abstract

A medical device system includes a medical device comprising one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes; a memory configured to store a machine learning model and a plurality of training data sets; and processing circuitry in communication with the memory. The processing circuit is configured to apply the machine learning model to cardiac electrical data to determine a value of a metric of left ventricular (LV) dysfunction. A machine learning model is trained based on the plurality of training data sets. Each training dataset of the plurality of training datasets includes a training cardiac electrical dataset and information indicative of one or more values of a metric of the LV dysfunction corresponding to the training cardiac electrical dataset.

Description

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 373,865, filed on August 30, 2022, the entire contents of which are incorporated herein by reference. Technical Field

[0002] The present disclosure relates to medical devices and, more particularly, to systems including medical devices and methods of operating such systems to monitor a patient's condition. Background Art

[0003] Some types of medical devices may be used to monitor one or more physiological parameters of a patient, such as physiological parameters associated with cardiac function. Such medical devices may include or be part of a system that includes a sensor that detects signals associated with such physiological parameters, such as heart rate parameters. Values ​​determined based on such signals may be used to assist in detecting changes in a medical condition, assessing the efficacy of a therapy, or generally assessing patient health.

[0004] A medical device that monitors a physiological parameter related to a patient's medical condition may evaluate a value associated with the physiological parameter, such as to determine whether the value exceeds a threshold or has changed over time. A value that exceeds a threshold or has changed may indicate a change (e.g., a worsening) in the patient's medical condition or that a therapy administered to the patient is not effectively controlling the patient's medical condition. Summary of the invention

[0005] In general, the present disclosure relates to techniques for determining a measure of left ventricular (LV) dysfunction, such as LV ejection fraction (LVEF), based on an electrocardiogram (ECG) or electrogram (EGM) sensed by a device such as an insertable cardiac monitor (ICM) or another implantable medical device (IMD). LV dysfunction may be associated with heart failure (HF). In some examples, processing circuitry of a system implementing the techniques of the present disclosure may further determine a HF state based on the measure of LV dysfunction, such as determining a likelihood of future worsening HF.

[0006] According to the technology of the present disclosure, an insertable cardiac monitor (ICM) or other implantable medical device (IMD) can collect cardiac signals indicative of cardiac electrical data such as ECG data or EGM data. The cardiac signals can include information relevant to monitoring a variety of patient conditions. For example, the cardiac signals can include reference points indicating events in the cardiac cycle. For example, a P wave indicates atrial depolarization, an R wave indicates ventricular depolarization, and a T wave indicates ventricular repolarization. The system can analyze the cardiac signals to identify these reference points, or monitor one or more patient conditions based on the identified reference points and / or other information. For example, the system can classify the cardiac signals as indicating one or more patient conditions such as atrial fibrillation (AF).

[0007] The disclosed techniques may be implemented by a system including one or more IMDs that are implanted in a patient for months or years while autonomously and continuously collecting biometric data and performing a large number of operations on the data per second to enable the system herein to analyze the biometric data. The use of the disclosed techniques with an IMD may be advantageous when a physician is unable to continuously accompany the patient to assess physiological parameters and / or when performing operations on the data described herein is deemed physically impossible by the physician.

[0008] In some examples, the techniques and systems of the present invention may use one or more models (e.g., probabilistic and / or machine learning models) to more accurately analyze biometric data to identify the value of a metric, monitor a patient condition, or predict a future state of a patient condition. In some examples, the model is trained using a set of training instances, where one or more of the training instances include data indicating a relationship between the biometric data and a likelihood value. Because the model is configured based on potentially thousands or millions of training instances, the model may reduce the amount of error in recursive predictions performed by the system. Reducing errors using the techniques of the present disclosure may provide one or more technical and clinical advantages, such as increasing the likelihood that the system will accurately provide information based on analyzing biometric data.

[0009] A machine learning model can provide more accurate output when the model is trained with a larger training dataset than when the model is trained with a smaller training dataset. This is because a larger training dataset provides a greater number of training instances than a smaller training dataset, which means that a model trained with a larger training dataset generates output based on a greater amount of knowledge than a model trained with a smaller training dataset. When a machine learning model is trained using supervised learning, the training data is labeled so that the model learns the association between features of the data and the labels. This means that the accuracy of a machine learning model trained using supervised learning can be limited when the amount of labeled training data available is limited.

[0010] In some examples, the health monitoring system may apply a machine learning model to cardiac electrical data generated by a medical device based on a cardiac signal sensed by the medical device from a patient to determine a value of a measure of LV dysfunction. When the health monitoring system applies the machine learning model to the cardiac electrical data, the machine learning model may process the cardiac electrical data to determine a value of the measure of LV dysfunction. In some examples, the machine learning model may output a determined value of the measure of LV dysfunction. In some examples, the machine learning model may output a confidence that the measure of LV dysfunction is below a threshold measure of LV dysfunction and / or a confidence that the measure of LV dysfunction is not below a threshold measure of LV dysfunction.

[0011] One or more characteristics of the cardiac electrical data set may indicate a value of a measure of LV dysfunction corresponding to the cardiac electrical data set. For example, the cardiac electrical data may indicate heart rate, heart rate variability, one or more arrhythmias, characteristics corresponding to R waves, P waves, T waves, or other cardiac events, or any combination thereof. By applying a machine learning model to the cardiac electrical data including these characteristics, the health monitoring system may be configured to cause the machine learning model to output a value of a measure of LV dysfunction. It may be beneficial to use a machine learning model to generate a value of a measure of LV dysfunction based on the cardiac electrical data so that the health monitoring system can determine a measure of LV dysfunction based on the cardiac electrical data collected by the medical device.

[0012] In some examples, multiple training data sets may be used to train a machine learning model applied by a health monitoring system. Each of the multiple training data sets may include a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set. That is, each of the one or more values ​​of the measure of LV dysfunction indicated by the training data set may be associated with a time related to the training cardiac electrical data set. This means that the training data set may be labeled with one or more values ​​of the measure of LV dysfunction at a time related to a characteristic of the training cardiac electrical data. When training the machine learning model, the model may learn to recognize patterns corresponding to the characteristics of the cardiac electrical data and the measure of LV dysfunction.

[0013] The techniques of the present disclosure may provide one or more advantages. For example, by determining the value of a measure of LV dysfunction using a machine learning model trained with a training data set that includes both cardiac electrical data and one or more values ​​of a measure of LV dysfunction corresponding to the cardiac electrical data, the system may more accurately determine the value of the measure of LV dysfunction compared to a system that does not use a machine learning model trained with cardiac electrical data labeled with values ​​of the measure of LV dysfunction. The machine learning model may also be evaluated on a patient-by-patient basis. For example, the system may determine whether the machine learning model is adequately performing in determining the value of the measure of LV dysfunction based on cardiac electrical data collected from a particular patient, and update the model if the performance is inadequate. Therefore, the machine learning model may perform better for individual patients compared to a system that does not update the machine learning model based on performance on a patient-by-patient basis.

[0014] In one example, a medical device system includes: a medical device including one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes; a memory configured to store a machine learning model and a plurality of training data sets; and a processing circuit in communication with the memory. The processing circuit is configured to apply the machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set. The processing circuit is also configured to output the determined value of the measure of LV dysfunction to a computing device.

[0015] In another example, a method of operating a medical device system including a medical device, the medical device including one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes, includes applying, by processing circuitry of the medical device system, a machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set, wherein the processing circuitry is in communication with a memory of the medical device system, the memory being configured to store the machine learning model and the plurality of training data sets. The method also includes outputting, by the processing circuitry, the determined value of the measure of LV dysfunction to a computing device.

[0016] In another example, a non-transitory computer-readable storage medium includes program instructions that, when executed by processing circuitry of a medical device system that includes one or more electrodes and is configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes, causes the processing circuitry to apply a machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set, wherein the processing circuitry is in communication with a memory of the medical device system that is configured to store the machine learning model and the plurality of training data sets. The program instructions also cause the processing circuitry to output the determined value of the measure of LV dysfunction to a computing device.

[0017] The present disclosure is intended to provide an overview of the subject matter described in the present disclosure. The present disclosure is not intended to provide an exclusive or exhaustive explanation of the systems, devices, and methods described in detail in the following figures and description. Further details of one or more examples of the present disclosure are set forth in the following figures and description. Other features, objectives, and advantages will be apparent from the description and drawings, as well as from the claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a block diagram illustrating an example medical device system including an implantable medical device (IMD) in conjunction with a patient in accordance with one or more techniques of the present disclosure.

[0019] Figure 2A is a perspective view showing an IMD that may be an insertable cardiac monitor (ICM) according to one or more techniques of the present disclosure. Figure 1 Example configuration of an IMD.

[0020] Figure 2B is a perspective view showing another IMD according to one or more techniques of the present disclosure, which may be a ICM from Figure 1 Another example configuration of an IMD.

[0021] Figure 3 is a diagram showing one or more techniques according to the present disclosure Figure 1 Block diagram of an example configuration of an IMD.

[0022] Figure 4 is a block diagram showing an example configuration of a computing device according to one or more techniques of the present disclosure, which may correspond to Figure 1 Any computing device (or two computing devices operating in coordination) of the computing devices.

[0023] Figure 5 is a diagram showing one or more techniques according to the present disclosure Figure 1 Block diagram of the operational perspective of a health monitoring system (HMS).

[0024] Figure 6 is a conceptual diagram illustrating an example machine learning model configured to output information corresponding to cardiac electrical data in accordance with one or more techniques of this disclosure.

[0025] Figure 7 is a block diagram illustrating an example of a machine learning model trained using supervised learning and / or reinforcement learning in accordance with one or more techniques of this disclosure.

[0026] Figure 8is a flow chart illustrating an example method for determining whether to use a machine learning model to process cardiac electrical data collected from a patient in accordance with one or more techniques of this disclosure.

[0027] Fig. 9 is a flow chart illustrating an example method for applying a machine learning model to determine a value for a measure of left ventricular (LV) dysfunction in accordance with one or more techniques of the present disclosure. DETAILED DESCRIPTION

[0028] The device including one or more physiological sensors includes at least one channel of electrodes for sensing cardiac electrical data such as electrocardiogram (ECG) data and / or electrogram (EGM) data. The processing circuit of the device or another device communicating with the sensing device determines a measure of left ventricular (LV) dysfunction based on cardiac electrical data, such as LV ejection fraction (LVEF). In some examples, the device may be an implantable medical device (IMD). The IMD may include a housing configured for subcutaneous implantation, on which one or more electrodes are positioned, such as an insertable cardiac monitor (ICM). In some instances, the IMD may be a leadless IMD. In other examples, the device may be one or more other implantable or external devices. Examples of one or more other implantable or external devices may include an implantable multi-channel cardiac pacemaker, an implantable cardioverter-defibrillator (ICD), an implantable pulse generator (IPG), a leadless (e.g., intracardiac) pacemaker, an extravascular pacemaker and / or ICD, or other IMD, or a combination of such IMDs, an external monitor, a drug pump, or a smart watch, or other consumer devices with physiological monitoring capabilities.

[0029] In any such examples, the processing circuitry may transmit the value of the patient's LV dysfunction metric to a remote computer, receive from the remote computer instructions for performing a medical intervention based on the patient's heart failure state, and transmit the instructions for performing the medical intervention to the user interface. Such instructions for performing the medical intervention may include at least one of a change in medication selection, a change in medication dosage, an instruction to schedule a visit with a clinician, or an instruction for the patient to seek medical attention. In this way, the diagnosis and / or treatment of the patient for the condition may be modified as needed between clinic visits, which may help avoid adverse medical events, such as symptom recurrence, acute heart failure (HF) decompensation, or hospitalization.

[0030] In some examples, the techniques described herein can enable identification of changes in a patient's state before the changes lead to symptoms and / or progression or development of the patient's condition or one or more additional medical conditions. Thus, the techniques described herein can help enable determination of the likelihood that a patient will experience an adverse medical event, which can help clinicians prescribe personalized treatments to help avoid hospitalizations, improve clinical outcomes, and / or reduce the economic burden on the healthcare system.

[0031] The present disclosure describes various systems, devices, and techniques for analyzing biometric data collected by a medical device to generate information related to a patient's health. In some examples, one or more patient conditions (such as HF, arrhythmia) can be identified or the future occurrence of one or more patient conditions can be predicted based on analyzing cardiac electrical data such as an ECG or EGM. The system can use a machine learning model to analyze cardiac data to generate information related to the patient's condition, such as one or more values ​​of a measure of LV dysfunction. In some cases, the system can train the machine learning model using a large training data set including one or both of an unlabeled training data set and a labeled training data set.

[0032] In machine learning, "labeled" training data refers to training data to which is attached information that identifies one or more aspects of the data. For example, cardiac electrical data such as ECG data and EGM data may include labels that identify P waves, R waves, and T waves. In another example, the cardiac electrical data may include labels that indicate parameters such as heart rate, heart rate variability, or characteristics such as arrhythmias. In another example, the cardiac electrical data may be labeled with one or more values ​​of a measure of LV dysfunction such as LVEF. The label that includes the value of the measure of LV dysfunction may indicate the time of the value of the measure of LV associated with one or more times of the cardiac electrical data. Other example labels may include patient information such as age, gender, pre-existing conditions, or any combination thereof. "Unlabeled" training data refers to training data that does not include information that identifies various aspects of the data. An unlabeled cardiac electrical training data set may include cardiac electrical data that includes features such as P waves, R waves, and T waves without any labels that identify such features or any other characteristics of the data set.

[0033] The system can be configured to train a machine learning model based on multiple training data sets. Each training data set can include a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set. That is, the training cardiac electrical data set can include one or more labels indicating the values ​​of the measure of LV dysfunction. The training cardiac electrical data set can include one or more labels indicating characteristics of the cardiac electrical data, but this is not required. The system can train a machine learning model based on multiple training data sets so that the machine learning model learns patterns corresponding to the cardiac electrical data and the measure of LV dysfunction.

[0034] LVEF is a clinically actionable measure of cardiac function that can be measured in a clinical setting using echocardiography, cardiac magnetic resonance imaging (MRI), or other modalities. In some examples, it may be beneficial to derive a measure of LV dysfunction, such as LVEF, based on cardiac electrical data, such as ECG data or EGM data. This is because an IMD can continuously record cardiac electrical data from a patient over an extended period of time, and deriving a measure of LV dysfunction, such as LVEF, based on cardiac electrical data can allow the system to track a measure of LV dysfunction over an extended period of time without requiring the patient to frequently visit a clinic for echocardiography and / or cardiac MRI measurements.

[0035] Using cardiac electrical data to monitor a measure of LV dysfunction may be beneficial because cardiac electrical data, such as ECG data and EGM data, has diagnostic relevance for measures indicative of LV dysfunction, such as LVEF. Using cardiac electrical data to monitor a measure of LV dysfunction may allow a system to achieve continuous monitoring of a measure of LV dysfunction over a longer period of time compared to systems using methods such as echocardiography and / or MRI. Using cardiac electrical data to monitor a measure of LV dysfunction may allow a system to more accurately monitor a measure of LV dysfunction in real time over an extended period of time and / or monitor a measure of LV dysfunction via cloud-based artificial intelligence (AI) compared to systems using echocardiography and / or MRI to measure a measure of LV dysfunction because cardiac electrical data may be more easily recorded by an IMD over an extended period of time compared to echocardiography and / or MRI.

[0036] Figure 1is a block diagram illustrating an example medical device system 2 including an IMD 10 in conjunction with a patient 4 in accordance with one or more techniques of the present disclosure. The IMD 10 is configured for continuous, long-term monitoring of the heart of the patient 4. For example, the IMD 10 is configured to sense a cardiac signal indicative of cardiac electrical data (such as ECG data, EGM data, or other cardiac electrical data), identify a fiducial point in the cardiac signal, identify one or more parameter values ​​(e.g., a value of a measure of LV dysfunction) based on the cardiac signal, detect a patient condition or arrhythmia (e.g., HF, ventricular fibrillation (VF), atrial fibrillation (AF), atrioventricular (AV) block) based on the cardiac signal, store the cardiac signal and information corresponding to the cardiac signal to a memory of the IMD 10, or any combination thereof. In some examples, the IMD 10 is configured to automatically sense the cardiac signal continuously or intermittently according to a predetermined schedule. In some examples, the IMD 10 is configured to sense the cardiac signal based on a user input. For example, the IMD 10 may sense the patient signal based on a user input indicating a symptom of the patient condition.

[0037] For example, IMD 10 can be a pacemaker or implantable cardioverter defibrillator that can be connected to intravascular or extravascular leads, or a pacemaker having a housing configured for implantation in the heart, which can be leadless. Some IMDs do not provide therapy, such as implantable patient monitors. IMD 10 can be such an IMD, such as the RevealLINQ 100 commercially available from Medtronic, Inc. of Minneapolis, Minnesota. TM or LINQ II TM An insertable cardiac monitor (ICM), which can be inserted subcutaneously. This type of IMD can facilitate relatively long-term and continuous monitoring of patients during normal daily activities, and can periodically transmit the collected data to a remote patient monitoring system, such as the Medtronic CareLink TM network.

[0038] IMD 10 can determine the value of patient parameter or metric, for example, based on the physiological signal sensed by IMD or the responsive therapy delivered by IMD. For example, patient parameter can include fluid level, heart rate, respiratory rate, patient activity, temperature, heart sound, oxygenation and R wave morphology characteristics. Other example patient metrics include cough, speech, posture, tissue perfusion, hematocrit, chest impedance, subcutaneous impedance, intracardiac impedance, heart rate variability (HRV), weight, blood pressure, sleep apnea burden (which can be derived from respiratory rate), ischemic burden, sleep duration, sleep quality, ventricular premature contraction (PVC) burden, occurrence, metric of LV dysfunction such as LVEF, and frequency or duration of arrhythmia or other events such as HF, VF, AF or AV block, and sensed cardiac interval (e.g., QT interval). Another example patient metric is ventricular rate during AF. Concentration or level of various substances in the patient, such as blood sugar, hematocrit, troponin and / or brain natriuretic peptide (BNP) level, can also be used as one or more patient metrics.

[0039] Medical device system 2 may include one or more sensors (e.g., for sensing the activity state of patient 4 and / or cardiac function of patient 4). The one or more sensors may collectively detect at least one first signal and at least one second signal that enable processing circuitry of medical device system 2 to determine whether the activity state of patient 4 satisfies at least one inactivity criterion based on such signals and to determine a value of at least one LV dysfunction metric such as LVEF. Although such processing circuitry may be housed within IMD 10 and / or within another medical device of medical device system 2, such as external device 12, for clarity, the processing circuitry may be described herein as a component of IMD 10.

[0040] In some examples, the one or more sensors may include one or more accelerometers or other sensors configured to detect at least one signal indicative of one or more aspects of an activity state of patient 4, such as activity level, posture, and / or breathing rate. In some instances, the one or more sensors may include a plurality of electrodes, which may be positioned on a housing of IMD 10. The plurality of electrodes may be configured to detect cardiac electrical signals, such as an ECG and / or an EGM.

[0041] In addition to IMD 10, medical device system 2 also includes one or more patient computing devices, such as patient computing devices 12A and 12B (collectively referred to as "patient computing devices 12"). Patient computing device 12 is configured for wireless communication with IMD 10. Computing device 12 can retrieve parameter data, ECG data and / or EGM data, LV dysfunction metric data, and other data from IMD 10. In some examples, computing device 12 takes the form of a personal computing device of patient 4. For example, computing device 12A can take the form of a smartphone of patient 4, and computing device 12B can take the form of a smart watch or other smart clothing of patient 4. In some examples, computing device 12 can be any computing device configured for wireless communication with IMD 10, such as a desktop computer, a laptop computer, or a tablet computer. Computing device 12 can be based on, for example or The Bluetooth Low Energy (BLE) protocol communicates with IMD 10 and each other. In some examples, only one of computing devices 12 (e.g., computing device 12A) is configured to communicate with IMD 10, such as due to executing software capable of communicating and interacting with the IMD (e.g., as part of a health monitoring application as described herein).

[0042] In some examples, computing device 12 (e.g. Figure 1 Wearable computing device 12B in the illustrated example may include electrodes and other sensors to sense physiological signals of patient 4, and may collect and store parameter data based on such signals. Computing device 12B may be incorporated into patient 4's clothing, such as within clothing, shoes, glasses, a watch or wristband, a hat, etc. In some examples, computing device 12B is a smartwatch or other accessory or peripheral device to smartphone computing device 12A.

[0043] One or more of computing devices 12 may be configured to communicate with various other devices or systems via network 16. For example, one or more of computing devices 12 may be configured to communicate with one or more computing systems (e.g., computing system 20) via network 16. Computing system 20 may be managed by the manufacturer of IMD 10 to, for example, provide cloud storage and analysis of collected data, maintenance and software services, or other networking functionality for their respective devices and their users. In some examples, computing system 20 may include the Medtronic Carelink TM Online or through Medtronic Carelink TM Network to implement.

[0044] like Figure 1As shown, the computing system 20 may include a processing circuit 22 and a memory 24. The processing circuit 22 may include a fixed function circuit and / or a programmable processing circuit. The processing circuit 22 may include any one or more of the following: a microprocessor, a controller, a digital signal processor (DSP), a graphics processing unit (GPU), a tensor processing unit (TPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or an equivalent discrete or analog logic circuit. In some examples, the processing circuit 22 may include multiple components (such as one or more microprocessors, one or more controllers, one or more DSPs, GPUs, TPUs, one or more ASICs, or any combination of one or more FPGAs) that may be physically located in one or more physical locations in one or more devices, as well as other discrete or integrated logic circuits. The computing system 20 may be configured as a cloud computing system.

[0045] Processing circuitry 22 may be capable of processing instructions stored in memory 24. In some examples, memory 24 includes computer-readable media including instructions that, when executed by processing circuitry 22, cause computing system 20 and processing circuitry 22 to perform the various functions attributed thereto herein. Figure 1 In the example shown, computing system 20 implements health monitoring system (HMS) 26. As will be described in more detail below, HMS 26 may generate information corresponding to one or more patient conditions of patient 4 based on physiological data collected by IMD 10. Memory 24 may include any volatile, nonvolatile, magnetic, optical, or electrical medium, such as random access memory (RAM), read-only memory (ROM), nonvolatile RAM (NVRAM), electrically erasable programmable ROM (EEPROM), ferroelectric RAM (FRAM), dynamic random access memory (DRAM), flash memory, or any other digital medium.

[0046] Computing device 12 may transmit data to computing system 20 via network 16, including data retrieved from IMD 10. The data may include: sensed data, such as values ​​of physiological parameters measured by IMD 10 and, in some cases, one or more computing devices in computing device 12; and other physiological signals or data recorded by IMD 10 and / or computing device 12. The data transmitted from computing device 12 to computing system 20 may include cardiac electrical data, such as ECG data and / or EGM data.

[0047] Network 16 may include one or more computing devices, such as one or more non-edge switches, routers, hubs, gateways, security devices (such as firewalls), intrusion detection, and / or intrusion prevention devices, servers, cellular base stations and nodes, wireless access points, bridges, cable modems, application accelerators, or other network devices. Network 16 may include one or more networks managed by a service provider and, therefore, may form part of a large-scale public network infrastructure (e.g., the Internet). Network 16 may provide computing devices and systems (such as Figure 1 The network 16 may include a dedicated network that provides a communication framework that allows the computing devices and systems to communicate with each other. Figure 1 The computing devices and systems illustrated in the example communicate with each other, but some of the data flows are isolated from devices outside the private network for security purposes. In some examples, Figure 1 Communications between the computing device and the system illustrated in are encrypted.

[0048] As will be described herein, IMD 10 may be configured to generate diagnostic information for patient 4, such as information indicating one or more values ​​of a metric of LV dysfunction. In some examples, IMD 10 may be configured to transmit such data to wireless access point 34 and / or computing device 12. Wireless access point 34 and / or computing device 12 may then communicate the retrieved data to computing system 20 via network 16.

[0049] In some cases, computing system 20 may be configured to provide a secure storage site for data that has been collected from IMD 10 and / or computing device 12. In some cases, computing system 20 may include a database that stores medical-related data and health-related data. For example, computing system 20 may include a cloud server or other remote server that stores data collected from IMD 10 and / or computing device 12. In some cases, computing system 20 may compile the data in a web page or other document for viewing by a trained professional such as clinician 40 via clinician computing device 38. Reference Figure 1 One or more aspects of the described example systems may be similar to those provided by Medtronic The network is implemented using general network technologies and functionality provided by the network.

[0050] In some examples, one or more of clinician computing devices 38 may be a tablet or other smart device located with the clinician through which the clinician may program, receive alerts from, and / or query IMD 10. For example, a clinician may access data collected by IMD 10 through clinician computing device 38, such as when patient 4 is between clinician visits, to check on the status of a medical condition. For example, computing system 20 may transmit data indicating one or more cardiac electrical data sets and / or one or more values ​​of a measure of LV dysfunction (e.g., determined by computing system 20, computing device 12, or other devices described herein) to clinician 40 via clinician computing device 38.

[0051] As will be described herein, processing circuitry of one or more devices of the medical device system 2 (e.g., computing device 12 and / or computing system 20 implementing HMS26) may be configured to determine information corresponding to cardiac electrical data (e.g., ECG data and / or EGM data) generated by the IMD 10. In some examples, the HMS26 may include a machine learning model that is configured to generate an output based on receiving as input the cardiac electrical data generated by the IMD 10. In some examples, the machine learning model may generate an output that includes information corresponding to the health of the patient 4. For example, the machine learning model may output one or more values ​​of a measure of LV dysfunction corresponding to the cardiac electrical data generated by the IMD 10. In some cases, the machine learning model may output information indicating the level of the measure of LV dysfunction that the patient 4 may experience in the future. In some examples, the machine learning model may be stored by a memory of the IMD 10, but this is not required. The machine learning model may be stored by a device separate from the IMD 10.

[0052] In some examples, computing system 20 may be configured to determine a long-term trend in a measure of LV dysfunction on a patient-specific basis. For example, computing system 20 may determine a historical estimate of a measure of LV dysfunction based on a value of a measure of LV dysfunction determined from cardiac electrical data collected by IMD 10 and actual values ​​of the measure of LV dysfunction measured via echocardiography and / or cardiac MRI. For example, computing system 20 may be configured to determine a long-term trend in a measure of LV dysfunction based on a value of a measure of LV dysfunction determined from cardiac electrical data collected by IMD 10 and actual values ​​of the measure of LV dysfunction obtained on a patient-specific basis using data corresponding to a given patient.

[0053] exist Figure 1In the example of , environment 28 includes one or more Internet of Things (IoT) devices, such as IoT device 30. IoT device 30 may include, for example, a so-called "smart" speaker, camera, television, light, lock, thermostat, appliance, actuator, controller, or any other smart home (or building) device. Figure 1 In the example of, IoT device 30 is a smart speaker and / or controller, which may include a display. In some examples, IoT device 30 includes a camera, or other sensors that can activate those sensors to collect data about patient 4, such as for evaluating the condition of patient 4.

[0054] Computing device 12 may be configured to communicate wirelessly with IoT device 30 to cause IoT device 30 to take actions described herein. In some examples, HMS 26 communicates with IoT device 30 via network 16 to cause IoT device 30 to take actions described herein. In some examples, IMD 10 is configured to communicate wirelessly with IoT device 30, for example to communicate data to computing system 20 via network 16. In some examples, IoT device 30 may be configured to provide some or all of the functionality attributed to computing device 12 herein.

[0055] The environment 28 includes a computing facility (e.g., a local network 32) through which the computing device 12, IoT device 30, and other devices within the environment 28 can communicate with the HMS 26 via the network 16, for example. For example, the environment 28 can be configured with wireless technologies, such as the Institute of Electrical and Electronics Engineers (IEEE) 802.11 wireless network, IEEE 802.15 ZigBee network, ultra-wideband protocol, near field communication, etc. The environment 28 may include one or more wireless access points, such as wireless access point 34, to provide support for wireless communication throughout the environment 28. Additionally or alternatively, for example, when the local network is not available, the computing device 12, IoT device 30, and other devices within the environment 28 can be configured to communicate with the network 16, such as the HMS 26, via a cellular base station 36 and a cellular network.

[0056] In some examples, computing device 12 and / or computing system 20 may implement one or more algorithms to determine information corresponding to patient 4 based on data received from IMD 10. In some examples, computing device 12 and / or computing system 20 may have greater processing power than IMD 10, enabling more complex analysis of the data. In some examples, computing device 12 and / or HMS 26 may apply the data to a probabilistic model, a machine learning model, or other artificial intelligence developed algorithm, for example, to determine information corresponding to patient 4 as described herein. Although described in the context of an example in which IMD 10 that senses a patient's cardiac activity may include an ICM, an example system including one or more implanted, wearable, or external devices of any type configured to sense a patient's physiological parameters may be configured to implement the techniques of the present disclosure.

[0057] In some examples, the IMD 10 includes one or more electrodes. The IMD 10 may generate cardiac electrical data (e.g., ECG data and / or EGM data) based on cardiac signals sensed via the one or more electrodes. In some examples, the cardiac electrical data may be based on cardiac electrical signals indicating one or more aspects of cardiac activity of the patient 4. For example, the cardiac electrical data may indicate one or more cardiac events, such as atrial depolarization (indicated by a P wave), ventricular depolarization (indicated by an R wave), and ventricular repolarization (indicated by a T wave). The cardiac data may represent a sequence of cardiac data points, such that any given data point in the sequence of cardiac data points corresponds to a magnitude of the cardiac signal at a point in time. Based on the time at which the cardiac event occurs, the cardiac data may indicate one or more parameters, such as heart rate and heart rate variability. In some examples, the IMD 10 may output the cardiac electrical signals to the computing system 20. In some cases, the computing system 20 may store the cardiac electrical signals in the memory 24.

[0058] In some examples, memory 24 is configured to store a machine learning model. HMS 26 may be configured to apply the machine learning model to cardiac electrical data (e.g., ECG data and / or EGM data) received from IMD 10 as input to generate an output. In some examples, the machine learning model is trained based on multiple training data sets. Memory 24 may be configured to store multiple training data sets. Each of the multiple training data sets may include a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set. In some examples, one or more values ​​of a measure of LV dysfunction corresponding to a training cardiac electrical data set of a training data set may each correspond to a time associated with one or more times of the training cardiac electrical data set. In some examples, a training cardiac electrical data set of a training data set may include one or more tags indicating characteristics of the training cardiac electrical data set, but this is not required.

[0059] In some examples, the cardiac electrical data received by computing system 20 from IMD 10 may include multiple cardiac electrical data sets. In some examples, the multiple cardiac electrical data sets may represent segments of a long sample of cardiac electrical data. In some examples, the multiple cardiac electrical data sets may each represent an individual data sample collected by IMD 10 at a different time or time range. In some examples, the multiple cardiac electrical data sets may each correspond to the same duration. In some examples, one or more cardiac electrical data sets may correspond to a duration that is different from the duration of one or more other cardiac electrical data sets in the multiple cardiac electrical data sets.

[0060] HMS26 may apply a machine learning model to each of a plurality of cardiac electrical data sets in order to determine a value of a measure of LV dysfunction corresponding to each of the plurality of cardiac electrical data sets. This may allow HMS26 to generate values ​​of a measure of LV dysfunction corresponding to different time points or time windows. By determining the value of a measure of LV dysfunction corresponding to each of a plurality of cardiac electrical data sets, HMS26 may track the measure of LV dysfunction over a certain period of time to determine one or more trends of the measure of LV dysfunction over the period of time. In some examples, HMS26 may determine a trend of the measure of LV dysfunction over a certain period of time based on one or more determined values ​​of the measure of LV dysfunction. For example, HMS26 may determine that the measure of LV dysfunction deteriorates over the period of time, improves over the period of time, or remains stable over the period of time.

[0061] IMD 10 may include a motion sensor, such as an accelerometer. The motion sensor of IMD 10 may generate motion data based on the motion of IMD 10. In some examples, IMD 10 may output the motion data to computing system 20 for analysis. In some examples, HMS26 may determine a motion value indicating the activity level of patient 4 based on the motion data received from IMD 10. The motion value corresponds to a value of a measure of LV dysfunction determined based on cardiac electrical data received from IMD 10. HMS26 may determine whether to output a warning based on the motion value and the determined value of the measure of LV dysfunction. For example, HMS26 may determine whether the value of the measure of LV dysfunction is lower than a threshold measure of LV dysfunction. HMS26 may determine whether the motion value is greater than a threshold motion value. HMS26 may determine whether to output a warning based on whether the value of the measure of LV dysfunction is lower than a threshold measure of LV dysfunction and whether the motion value is greater than a threshold motion value.

[0062] For example, when the measure of LV dysfunction is LVEF, it may be beneficial to output a warning when the patient's LVEF is low and the patient's activity level is high. In this example, HMS 26 may output a warning when the motion value is greater than a threshold motion value and when the value of the measure of LV dysfunction is below the threshold measure of LV dysfunction. In some examples, HMS 26 may not output a warning when the motion value is not greater than the threshold motion value or when the value of the measure of LV dysfunction is not below the threshold measure of LV dysfunction.

[0063] HMS 26 may train a machine learning model based on multiple training data sets. By training the machine learning model based on multiple training data sets, the processing circuit is configured to enable the machine learning model to identify one or more patterns corresponding to a measure of LV dysfunction and one or more characteristics of the cardiac electrical data. Since each of the multiple training data sets may include a cardiac electrical data set and one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set, each training data set includes cardiac electrical data labeled with values ​​of the measure of LV dysfunction. Based on the characteristics of the cardiac electrical data and the labels indicating the values ​​of the measure of LV dysfunction in each training data set, HMS 26 may train the machine learning model to recognize patterns corresponding to the cardiac electrical data and the measure of LV dysfunction.

[0064] For example, once the machine learning model has been trained, HMS26 may apply the machine learning model to a cardiac electrical data set indicating one or more characteristics and / or parameters (e.g., P wave, T wave, R wave, heart rate, heart rate variability) to determine the value of the measure of LV dysfunction. HMS26 may apply the machine learning model to process the cardiac electrical data set to determine the value of the measure of LV dysfunction based on one or more characteristics of the cardiac electrical data set and one or more learning patterns between the characteristics of the cardiac electrical signal and the measure of LV dysfunction. In some examples, the machine learning model may output an estimated true value of the measure of LV dysfunction. When the measure of LV dysfunction includes LVEF, the true value of the measure of LV dysfunction may include the fraction of the amount of blood pumped from the left ventricle to the total amount of blood in the left ventricle. In some examples, the machine learning model may output a confidence value indicating that the true value of the measure of LV dysfunction is accurate. In some examples, the machine learning model may output a confidence that the value of the measure of LV dysfunction corresponding to the cardiac electrical data set is lower than a threshold value of the measure of LV dysfunction. In some examples, the machine learning model may output a confidence level that a value of a measure of LV dysfunction corresponding to the cardiac electrical dataset is not below a threshold value for the measure of LV dysfunction.

[0065] In some examples, HMS26 may mark each of the multiple training data sets stored in memory 24. In some examples, in order to mark each of the multiple training data sets, HMS26 may identify one or more characteristics of the training cardiac electrical data set in the training data set for each of the multiple training data sets. For example, HMS26 may identify one or more P waves, R waves, T waves, arrhythmic events, other cardiac electrical features, or any combination thereof. Additionally or alternatively, HMS26 may identify one or more parameters corresponding to the training cardiac electrical data set, such as heart rate, heart rate variability, pulse transit time (PTT), or any combination thereof. HMS26 may use one or more characteristics of the training cardiac electrical data set to mark the training cardiac electrical data set of each of the multiple training data sets.

[0066] In some examples, to label each of the plurality of training data sets, HMS 26 may identify the time corresponding to each of the one or more values ​​of the measure of LV dysfunction in the information indicating the one or more values ​​of the measure of LV dysfunction of the training data set. In some examples, the time corresponding to each of the one or more values ​​of the measure of LV dysfunction corresponds to the time at which the value of the measure of LV dysfunction was measured. HMS 26 may associate the time corresponding to each of the one or more values ​​of the measure of LV dysfunction with the time of the training cardiac electrical data set. That is, HMS 26 may label each of the plurality of training data sets with the one or more times corresponding to the cardiac electrical data and the time corresponding to each of the one or more values ​​of the measure of LV dysfunction. This means that when each of the plurality of training data sets is labeled, the labeled training data indicates the time of the value of the measure of LV dysfunction measured relative to the time of the cardiac electrical data.

[0067] The techniques described herein are not limited to examples in which IMD 10 includes an ICM. In some examples, Figure 1 The IMD 10 may include any type of external or implanted medical device configured to collect cardiac electrical data (such as ECG data or EGM data). For example, the IMD 10 may include an implantable cardioverter-defibrillator (ICD), a pacemaker, a cardiac resynchronization therapy pacemaker (CRT-P), a cardiac resynchronization therapy defibrillator (CRT-D), or an external cardiac signal sensing device.

[0068] Figure 2A is a perspective view showing an IMD 10A, which may be an ICM, according to one or more techniques of the present disclosure. Figure 1 Example configuration of an IMD 10. Figure 2AIn the example shown, IMD 10A can be embodied as a monitoring device having a housing 112, a proximal electrode 116A, and a distal electrode 116B. Housing 112 can also include a first major surface 114, a second major surface 118, a proximal end 120, and a distal end 122. Housing 112 encloses electronic circuitry located within IMD 10A and protects the circuitry contained therein from bodily fluids. Housing 112 can be hermetically sealed and configured for subcutaneous implantation. Electrical feedthroughs provide electrical connections for electrodes 116A and 116B.

[0069] exist Figure 2A In the example shown, IMD 10A is defined by a length L, a width W, and a thickness or depth D, and is in the form of an elongated rectangular prism, wherein length L is much greater than width W, which in turn is greater than depth D. In one example, the geometry of IMD 10A, particularly width W being greater than depth D, is selected to allow IMD 10A to be inserted beneath the skin of a patient using a minimally invasive procedure and maintained in a desired orientation during insertion. For example, Figure 2A The device shown includes radial asymmetry (particularly a rectangular shape) along the longitudinal axis, which maintains the device in the correct orientation after insertion. For example, the spacing between the proximal electrode 116A and the distal electrode 116B can be in the range of 5 millimeters (mm) to 55mm, 30mm to 55mm, 35mm to 55mm, and 40mm to 55mm, and can be any range or single spacing of 5mm to 60mm. In addition, the length L of the IMD 10A can be in the range of 30mm to about 70mm. In other examples, the length L can be in the range of 5mm to 60mm, 40mm to 60mm, 45mm to 60mm, and can be any length or length range between about 30mm and about 70mm. In addition, the width W of the first major surface 114 can be in the range of 3mm to 15mm, 3mm to 10mm, or 5mm to 15mm, and can be any single width or width range between 3mm to 15mm. The thickness of the depth D of the IMD 10A may be in the range of 2 mm to 15 mm, 2 mm to 9 mm, 2 mm to 5 mm, 5 mm to 15 mm, and may be any single depth or depth range between 2 mm and 15 mm. In addition, the IMD 10A according to the examples of the present disclosure has a geometry and size designed for ease of implantation and patient comfort. The examples of the IMD 10A described in the present disclosure may have a volume of three cubic centimeters (cm) or less, a volume of 1.5 cubic centimeters or less, or any volume between three cubic centimeters and 1.5 cubic centimeters.

[0070] exist Figure 2AIn the example shown, once inserted into the patient, the first major surface 114 faces outward, toward the patient's skin, and the second major surface 118 is located opposite the first major surface 114. Figure 2A In the example shown, proximal end 120 and distal end 122 are rounded to reduce discomfort and irritation to surrounding tissue once inserted under the patient's skin. IMD 10A, including instruments and methods for inserting IMD 10A, is described, for example, in U.S. Pat. No. 11,311,312, which is incorporated herein by reference in its entirety.

[0071] Proximal electrode 116A is located at or near proximal end 120, and distal electrode 116B is located at or near distal end 122. Proximal electrode 116A and distal electrode 116B are used to sense cardiac electrical signals (e.g., ECG signals and / or EGM signals) and can measure thoracic interstitial impedance outside the thorax submuscularly or subcutaneously. The cardiac electrical signals and impedance measurements can be stored in a memory of IMD 10A, and the data can be transmitted to another device, which can be another implanted device or an external device, such as computing device 12, via integrated antenna 126A. In some examples, electrodes 116A and 116B can be used additionally or alternatively to sense any biopotential signal of interest from any implant location, which can be, for example, an EGM, electroencephalogram (EEG), electromyogram (EMG), or neural signal. Housing 112 can accommodate Figure 3 The circuit of IMD 10 is shown in FIG.

[0072] exist Figure 2A In the example shown, the proximal electrode 116A is located at or near the proximal end 120, and the distal electrode 116B is located at or near the distal end 122. In this example, the distal electrode 116B is not limited to a flat, outwardly facing surface, but can extend from the first major surface 114 around the rounded edge 124 and / or the end surface 126 to the second major surface 118, so that the electrode 116B has a three-dimensional curved configuration. In some examples, the electrode 116B is a non-insulated portion of the metal (e.g., titanium) portion of the housing 112.

[0073] exist Figure 2A In the example shown, the proximal electrode 116A is located on the first major surface 114 and is substantially flat and faces outward. However, in other examples, the proximal electrode 116A can utilize a three-dimensional curved configuration of the distal electrode 116B, thereby providing a three-dimensional proximal electrode (not shown in this example). Similarly, in other examples, the distal electrode 116B can utilize a substantially flat and outward facing electrode located on the first major surface 114, similar to the electrode shown with respect to the proximal electrode 116A.

[0074] Various electrode configurations allow for configurations in which the proximal electrode 116A and the distal electrode 116B are located on the first major surface 114 and the second major surface 118. In other configurations, such as Figure 2A In the configuration shown, only one of proximal electrode 116A and distal electrode 116B is located on both first major surface 114 and second major surface 118, while in other configurations, proximal electrode 116A and distal electrode 116B are both located on one of first major surface 114 or second major surface 118 (i.e., proximal electrode 116A is located on first major surface 114 and distal electrode 116B is located on second major surface 118). In another example, IMD 10A may include electrodes on both first major surface 114 and second major surface 118 at or near the proximal and distal ends of the device, such that a total of four electrodes are included on IMD 10A. Electrodes 116A and 116B may be formed of a variety of different types of biocompatible conductive materials (e.g., stainless steel, titanium, platinum, iridium, or alloys thereof), and may utilize one or more coatings, such as titanium nitride or fractal titanium nitride.

[0075] exist Figure 2A In the example shown, the proximal end 120 includes a head assembly 128 that includes one or more of the proximal electrode 116A, the integrated antenna 126A, the anti-migration protrusion 132, and / or the suture hole 134. The integrated antenna 126A is located on the same major surface (e.g., the first major surface 114) as the proximal electrode 116A and is also included as part of the head assembly 128. The integrated antenna 126A allows the IMD 10A to transmit and / or receive data. In other examples, the integrated antenna 126A may be formed on a major surface opposite the proximal electrode 116A, or may be incorporated into the housing 112 of the IMD 10A. Figure 2A In the example shown, the anti-migration protrusion 132 is located adjacent to the integrated antenna 126A and protrudes away from the first major surface 114 to prevent longitudinal movement of the device. Figure 2A In the example shown, the anti-migration protrusions 132 include a plurality (e.g., nine) of small bumps or protrusions extending away from the first major surface 114. As discussed above, in other examples, the anti-migration protrusions 132 may be located on the major surface opposite the proximal electrode 116A and / or the integrated antenna 126A. Figure 2A In the example shown, head assembly 128 includes suture holes 134 that provide another means of securing IMD 10A to the patient to prevent movement after insertion. In the example shown, suture holes 134 are located near proximal electrode 116A. In one example, head assembly 128 is a molded head assembly made of a polymer or plastic material that can be integrated or separated from the main portion of IMD 10A.

[0076] In some examples, IMD 10A may be configured to collect cardiac electrical signals, such as ECG signals or EGM signals, via electrodes 116A, 116B. In some examples, the cardiac electrical signals collected by IMD 10A may include one or more characteristics and / or indicate one or more patient parameters. IMD 10A may be configured to process the cardiac electrical signals, pre-process the cardiac electrical signals, output the cardiac electrical signals, store the cardiac electrical signals in a memory, or any combination thereof.

[0077] Figure 2B is a perspective view showing another IMD 10B according to one or more techniques of the present disclosure, which may be a ICM from Figure 1 Another example configuration of IMD 10 is shown. Figure 2B The IMD 10B can be basically similar to Figure 2A The IMD 10A is configured, the differences between which are discussed in this article.

[0078] IMD 10B may include a leadless, subcutaneously implantable monitoring device, such as an ICM. IMD 10B includes a housing having a base 140 and an insulating cover 142. Proximal electrode 116C and distal electrode 116D may be formed or placed on an outer surface of insulating cover 142. For example, regarding Figure 3 The various circuits and components of the described IMD 10B may be formed or placed on an inner surface of insulating cover 142 or within base 140. In some examples, a battery or other power source for IMD 10B may be included within base 140. In the example shown, antenna 126B is formed or placed on an outer surface of insulating cover 142, but in some examples, may be formed or placed on an inner surface. In some examples, insulating cover 142 may be positioned above open base 140 so that base 140 and insulating cover 142 enclose circuits and other components and protect these circuits and other components from fluids such as body fluids. The housing including base 140 and insulating cover 142 may be hermetically sealed and configured for subcutaneous implantation.

[0079] Circuits and components may be formed on the inner side of insulating cover 142, such as by using flip chip technology. Insulating cover 142 may be flipped onto base 140. When flipped and placed onto base 140, components of IMD 10B formed on the inner side of insulating cover 142 may be positioned in gap 144 defined by base 140. Electrodes 116C and 116D and antenna 126B may be electrically connected to the circuit formed on the inner side of insulating cover 142 through one or more through holes (not shown) formed through insulating cover 142. Insulating cover 142 may be formed of sapphire (i.e., corundum), glass, poly-p-xylene, and / or any other suitable insulating material. Base 140 may be formed of titanium or any other suitable material (e.g., biocompatible material). Electrodes 116C and 116D may be formed of any one of stainless steel, titanium, platinum, iridium, or alloys thereof. Among other things, electrodes 116C and 116D may be coated with a material such as titanium nitride or fractal titanium nitride, although other suitable materials and coatings for such electrodes may also be used.

[0080] exist Figure 2B In the example shown, the housing of the IMD 10B defines a length L, a width W, and a thickness or depth D, and is in the form of an elongated rectangular prism, wherein the length L is much greater than the width W, which in turn is greater than the depth D, with Figure 2A 10A. For example, the spacing between the proximal electrode 116C and the distal electrode 116D may be in the range of 5 mm to 50 mm, 30 mm to 50 mm, 35 mm to 45 mm, and may be any single spacing or spacing range from 5 mm to 50 mm, such as about 40 mm. In addition, the IMD 10B may have a length L ranging from 5 mm to about 70 mm. In other examples, the length L may be in the range of 30 mm to 70 mm, 40 mm to 60 mm, 45 mm to 55 mm, and may be any single length or length range from 5 mm to 50 mm, such as about 45 mm. In addition, the width W may be in the range of 3 mm to 15 mm, 5 mm to 15 mm, 5 mm to 10 mm, and may be any single width or width range from 3 mm to 15 mm, such as about 8 mm. The thickness or depth D of the IMD 10B may be in the range of 2 mm to 15 mm, 5 mm to 15 mm, or 3 mm to 5 mm, and may be any single depth or depth range between 2 mm and 15 mm, such as approximately 4 mm. The IMD 10B may have a volume of three cubic centimeters (cm) or less, or 1.5 cubic centimeters or less (such as approximately 1.4 cubic centimeters).

[0081] exist Figure 2B In the example shown, once inserted subcutaneously into the patient, the outer surface of the insulating cover 142 faces outward, toward the patient's skin. Figure 2B As shown, proximal end 146 and distal end 148 are rounded to reduce discomfort and irritation to surrounding tissue once inserted beneath the patient's skin. Additionally, the edges of IMD 10B can be rounded.

[0082] In some examples, IMD 10B may be configured to collect cardiac electrical signals, such as ECG signals or EGM signals, via electrodes 116C, 116D. In some examples, the cardiac electrical signals collected by IMD 10B may include one or more characteristics and / or indicate one or more patient parameters. IMD 10B may be configured to process the cardiac electrical signals, pre-process the cardiac electrical signals, output the cardiac electrical signals, store the cardiac electrical signals in a memory, or any combination thereof.

[0083] Figure 3 is a diagram showing one or more techniques according to the present disclosure Figure 1 A block diagram of an example configuration of an IMD 10 is shown. Figure 3 As shown, IMD 10 includes processing circuitry 150 , memory 152 , sensing circuitry 154 coupled to electrodes 116A and 116B (hereinafter “electrodes 116A, 116B”) and one or more sensors 158 , and communication circuitry 160 .

[0084] Processing circuit 150 may include fixed function circuits and / or programmable processing circuits. Processing circuit 150 may include any one or more of a microprocessor, a controller, a GPU, a TPU, a DSP, an ASIC, an FPGA, or an equivalent discrete or analog logic circuit. In some examples, processing circuit 150 may include multiple components (such as any combination of one or more microprocessors, one or more controllers, one or more GPUs, one or more TPUs, one or more DSPs, one or more ASICs, or one or more FPGAs), as well as other discrete or integrated logic circuits. The functions attributed to processing circuit 150 herein may be embodied as software, firmware, hardware, or any combination thereof. In some examples, memory 152 includes computer-readable instructions that, when executed by processing circuit 150, cause IMD 10 and processing circuit 150 to perform various functions attributed to IMD 10 and processing circuit 150 herein. Memory 152 may include any volatile, non-volatile, magnetic, optical, or electrical medium, such as RAM, ROM, NVRAM, EEPROM, flash memory, or any other digital medium.

[0085] Sensing circuit 154 may sense cardiac signals via electrodes 116A, 116B and measure impedance, such as the impedance of tissue proximal to IMD 10. The measured impedance may vary based on respiration, cardiac pulse or flow, and the degree of perfusion or edema. Processing circuit 150 may determine patient metrics related to respiration, fluid retention, cardiac pulse or flow, perfusion, and / or edema based on the measured impedance. In some examples, processing circuit 150 may identify features of the sensed cardiac signals, such as heart rate, heart rate variability, T wave alternans, intrabeat intervals (e.g., QT intervals), and / or morphological features to detect the onset of an arrhythmia in patient 4.

[0086] In some examples, the IMD 10 includes one or more sensors 158, such as one or more accelerometers, gyroscopes, microphones, optical sensors, temperature sensors, pressure sensors, and / or chemical sensors. In some examples, the sensing circuit 154 may include one or more filters and amplifiers for filtering and amplifying signals received from one or more of the electrodes 116A, 116B and / or the sensors 158. In some examples, the sensing circuit 154 and / or the processing circuit 150 may include a rectifier, a filter and / or an amplifier, a sensing amplifier, a comparator and / or an analog-to-digital converter. The processing circuit 150 may determine parameter data 182 (e.g., a physiological parameter value of the patient 4) based on the signals from the sensors 158, which may be stored in the memory 152 as data 180. The patient parameters determined based on the signals from the sensors 158 may include intravascular fluid levels, interstitial fluid levels, oxygen saturation, glucose levels, stress hormone levels, heart sounds, body movements, activity intensity, sleep duration, sleep quality, body posture, or blood pressure.

[0087] In addition to data 180, memory 152 may store applications 170 that may be executed by processing circuitry 150. Applications 170 may include data processing application 172. Processing circuitry 150 may execute data processing application 172 to process parameter data 182 and cardiac electrical data 184. In some examples, data processing application 172 may identify information corresponding to parameter data 182 and / or cardiac electrical data 184. For example, data processing application 172 may identify the heart rate of patient 4 based on cardiac electrical data 184 by identifying the rate at which R waves appear in cardiac electrical data 184.

[0088] Data 180 may include cardiac electrical data 184 sensed by IMD 10 via electrodes 116A, 116B. In some examples, cardiac electrical data 184 may represent cardiac data, such as ECG data or EGM data indicating cardiac activity of patient 4. In some examples, when IMD 10 is implanted under the skin of patient 4, cardiac electrical data 184 may indicate cardiac activity of patient 4 over a long period of time (e.g., weeks or months) collected continuously via electrodes 116A, 116B. In some examples, when IMD 10 is implanted under the skin of patient 4, cardiac electrical data 184 may include multiple cardiac data sets each collected via electrodes 116A, 116B. Cardiac electrical data 184 may include data indicating cardiac signals, heart rate information, RR interval information, morphology information, or other information.

[0089] Processing circuit 150 may use communication circuit 160 to communicate parameter data 182 and cardiac electrical data 184 to one or more other computing devices, such as computing device 12 and / or computing system 20. Communication circuit 160 may include any suitable hardware, firmware, software, or any combination thereof for wirelessly communicating with another device. Communication circuit 160 may be configured to communicate via inductive coupling, electromagnetic coupling, near field communication (NFC), radio frequency (RF), Wi-Fi or other proprietary or non-proprietary wireless communication schemes to transmit and / or receive signals.

[0090] In some examples, memory 152 of IMD 10 is configured to store model 194 including machine learning model 196. IMD 10 may be configured to apply machine learning model 196 to cardiac electrical data 184 to determine a value of a measure of LV dysfunction corresponding to cardiac electrical data 184. In some examples, machine learning model 196 may be configured to Figure 1 The data 180 may be trained by computing system 20 of the IMD 10 and output to the IMD 10. The IMD 10 need not store models 194 including machine learning models 196. In some examples, the IMD 10 may output data 180 to one or more other devices (e.g., external device 12, computing system 20, clinician computing device 38, or any combination thereof) for processing.

[0091] Figure 4 1 is a block diagram illustrating an example configuration of computing device 12, which may correspond to either computing device 12A and 12B (or two computing devices operating in coordination), in accordance with one or more techniques of this disclosure. In some examples, computing device 12 takes the form of a smartphone, laptop computer, tablet computer, personal digital assistant (PDA), smart watch, or other wearable computing device. In some examples, IoT device 30 and / or clinician computing device 38 may be similar to Figure 4 1 is configured based on the configuration of computing device 12 shown in .

[0092] like Figure 4 As shown in the example of , the computing device 12 may be logically divided into a user space 202, a kernel space 204, and hardware 206. The hardware 206 may include one or more hardware components that provide an operating environment for components executing in the user space 202 and the kernel space 204. The user space 202 and the kernel space 204 may represent different sections or segments of memory, wherein the kernel space 204 provides higher privileges to processes and threads than the user space 202. For example, the kernel space 204 may include an operating system 220 that operates with higher privileges than components executing in the user space 202.

[0093] like Figure 4 As shown, hardware 206 includes processing circuitry 230, memory 232, one or more input devices 234, one or more output devices 236, one or more sensors 238, and communication circuitry 240. Although for purposes of illustration, Figure 4 1 is shown as a standalone device, but the computing device 12 may be any component or system including a processing circuit or other suitable computing environment for executing software instructions, and does not necessarily include, for example, Figure 4 One or more elements shown in .

[0094] Processing circuitry 230 is configured to implement functionality and / or processing instructions for execution within computing device 12. For example, processing circuitry 230 may be configured to receive and process instructions stored in memory 232 that provide functionality for components included in kernel space 204 and user space 202 to perform one or more operations in accordance with the techniques of this disclosure. Examples of processing circuitry 230 may include any one or more microprocessors, controllers, GPUs, TPUs, DSPs, ASICs, FPGAs, or equivalent discrete or integrated logic circuits.

[0095] Memory 232 may be configured to store information within computing device 12 for processing during operation of computing device 12. In some examples, memory 232 is described as a computer-readable storage medium. In some examples, memory 232 includes temporary memory or volatile memory. Examples of volatile memory include RAM, DRAM, static random access memory (SRAM), and other forms of volatile memory known in the art. In some examples, memory 232 also includes one or more memories configured for long-term storage of information, such as non-volatile storage elements. Examples of such non-volatile storage elements include magnetic hard disks, optical disks, floppy disks, flash memory, or various forms of electrically programmable memory (EPROM) or EEPROM memory. In some examples, memory 232 includes a storage device associated with a cloud.

[0096] One or more input devices 234 of computing device 12 may receive input, for example, from patient 4, clinician 40, or another user. Examples of input are tactile, audio, dynamic, and optical input. As examples, input device 234 may include a mouse, keyboard, voice response system, camera, button, control pad, microphone, presence-sensitive or touch-sensitive component (e.g., screen), or any other device for detecting input from a user or machine.

[0097] One or more output devices 236 of computing device 12 may generate output, for example, to patient 4 or another user. Examples of output are tactile output, haptic output, audio output, and visual output. Output devices 236 of computing device 12 may include a presence-sensitive screen, a sound card, a video graphics adapter card, a speaker, a cathode ray tube monitor, a liquid crystal display (LCD), a light emitting diode (LED), or any type of device for generating tactile, audio, and / or visual output.

[0098] One or more sensors 238 of computing device 12 may sense physiological parameters or signals of patient 4. Sensors 238 may include electrodes, accelerometers (e.g., 3-axis accelerometers), optical sensors, impedance sensors, temperature sensors, pressure sensors, heart sound sensors (e.g., microphones or accelerometers), and other sensors, as well as sensing circuits (e.g., including analog-to-digital converters (ADCs)), similar to those described above with respect to IMD 10 and Figure 3 Those described.

[0099] The communication circuitry 240 of the computing device 12 may communicate with other devices by transmitting and receiving data. The communication circuitry 240 may receive data from the IMD 10, such as receiving patient measurements and / or higher resolution diagnostic information from the communication circuitry in the IMD 10. The communication circuitry 240 may include a network interface card, such as an Ethernet card, an optical transceiver, a radio frequency transceiver, or any other type of device that can send and receive information. For example, the communication circuitry 160 may include a radio transceiver configured for use in accordance with a wireless communication protocol such as 3G, 4G, 5G, WiFi (e.g., 802.11 or 802.15 ZigBee), Or BLE standards or protocols to communicate.

[0100] like Figure 4 As shown, the health monitoring application 250 is executed in the user space 202 of the computing device 12. The health monitoring application 250 can be logically divided into a presentation layer 252, an application layer 254, and a data layer 256. The presentation layer 252 may include a user interface (UI) component 260 that generates and presents a user interface of the health monitoring application 250.

[0101] The data layer 256 may include parameter data 290 and cardiac electrical data 292, which may be received from the IMD 10 via the communication circuit 240 and stored in the memory 232 by the processing circuit 230. The application layer 254 may include, but is not limited to, a data analyzer 270 and a model configuration service 272. The data analyzer 270 may be configured to process the parameter data 290 and / or cardiac electrical data 292 generated by the IMD 10 to generate information corresponding to one or more patient conditions of the patient 4. The data analyzer 270 may determine information corresponding to the patient 4 based on applying the parameter data 290 and / or cardiac electrical data 292 as input to one or more models 294, which may include one or more probability models, machine learning models, algorithms, decision trees, and / or thresholds. In examples where the model 294 includes one or more machine learning models, the data analyzer 270 may apply a feature vector derived from the data to the model 294.

[0102] The model configuration service 272 may be configured to modify the model 294 based on feedback received from the HMS 26 indicating whether the parameters are accurate or updated, for example. In some examples, the model configuration service 272 may utilize a dataset from the patient 4 for supervised machine learning to further train the model included as part of the model 294. The model configuration service 272 or another component executed by the processing circuitry of the medical device system 2 may select a configuration of the model 294 based on the etiology data of the patient. In some examples, different models 294 customized for different patient cohorts may be used to select for the patient 4 based on such etiology data.

[0103] In some examples, model 294 includes a machine learning model that is configured to process cardiac electrical data 292 to determine a value of a measure of LV dysfunction. Computing device 12 may be configured to apply the machine learning model to cardiac electrical data 292 to determine a value of a measure of LV dysfunction corresponding to cardiac electrical data 292. In some examples, the machine learning model may be configured to process cardiac electrical data 292 to determine a value of a measure of LV dysfunction corresponding to cardiac electrical data 292. Figure 1 The computing system 20 and the output computing device 12 are trained. The computing device 12 need not store the model 294 including the machine learning model. In some examples, the computing device 12 may output the cardiac electrical data 292 to one or more other devices (e.g., the computing system 20, the clinician computing device 38, or any combination thereof) for processing and / or receiving one or more values ​​determined based on the processing of the cardiac electrical data 292.

[0104] Figure 5 is a block diagram illustrating an operational perspective of HMS 26 in accordance with one or more techniques of this disclosure. HMS 26 may be implemented in computing system 20, which may include hardware components embodied in one or more physical devices, such as processing circuitry 22, memory 24, and communication circuitry. Figure 5 Provides an operational perspective of the HMS 26 when hosted as a cloud-based platform. Figure 5 In the example of FIG. 1 , the components of HMS 26 are arranged according to multiple logical layers that implement the technology of the present disclosure. Each layer can be implemented by one or more modules including hardware, software, or a combination of hardware and software.

[0105] Computing devices such as computing device 12, IoT device 30, and clinician computing device 38 operate as clients that communicate with HMS 26 via interface layer 300. Computing devices typically execute client software applications, such as desktop applications, mobile applications, and web applications. Interface layer 300 represents a set of application programming interfaces (APIs) or protocol interfaces provided and supported by HMS 26 for client software applications. Interface layer 300 can be implemented with one or more web servers.

[0106] like Figure 5As shown, HMS26 also includes an application layer 302, which represents a collection of services 310 for implementing the functionality attributed to HMS26 in this article. The application layer 302 receives information from client applications, such as data from a computing device 12 or an IoT device 30 (some or all of these data may have been retrieved from the IMD 10), and further processes the information in response to the information according to one or more services in the service 310. The application layer 302 can be implemented as one or more discrete software services 310 executed on one or more application servers (e.g., physical or virtual machines). That is, the application server provides an operating environment for executing the service 310. In some examples, the functional interface layer 300 and the functionality of the application layer 302 as described above can be implemented at the same server. The service 310 can communicate via a logical service bus 312. The service bus 312 generally represents a logical interconnection or interface set that allows different services 310 to send messages to other services such as through a publish / subscribe communication model.

[0107] The data layer 304 of HMS 26 uses one or more data repositories to provide persistence for information in HMS 26. A data repository may generally be any data structure or software that stores and / or manages data. Examples of data repositories include, but are not limited to, relational databases, multidimensional databases, mappings, and hash tables, to name a few.

[0108] Services 310 may include data analyzer 330, model configuration service 332, and record management service 334. Figure 5 As shown, each service in service 310 is implemented in module form within HMS26. Although shown as a separate module for each service, in some examples, the functionality of two or more services can be combined into a single module or component. Each service in service 310 can be implemented in software, hardware, or a combination of hardware and software. In addition, service 310 can be implemented as a stand-alone device, a separate virtual machine or container, a process, a thread, or a software instruction generally used to execute on one or more physical processors. Record management service 334 can store received patient data as parameter data 350 and cardiac electrical data 352.

[0109] The data analyzer 330 may determine information corresponding to the cardiac activity of the patient 4 based on the cardiac electrical data 352 generated by the IMD 10 and, in some cases, other parameter data 350. In some examples, the data analyzer 330 may identify one or more features (e.g., R waves, T waves, P waves) in the cardiac electrical data 352 received from the IMD 10. In some examples, the data analyzer 330 may determine whether the cardiac electrical data 352 indicates a patient condition or an arrhythmia (such as HF, VF, AF, or AV block). The data analyzer 330 may determine information corresponding to the patient 4 based on applying the parameter data 350 and / or the cardiac electrical data 352 as input to the machine learning model 354. In some examples, the data analyzer 330 may apply a feature vector derived from the data to the machine learning model 354.

[0110] The machine learning model 354 may be developed by the model configuration service 332. Example machine learning techniques that may be used to generate the machine learning model 354 include various learning styles, such as supervised learning, unsupervised learning, and semi-supervised learning. Example types of algorithms include Bayesian algorithms, Markov models, Hawkes processes, clustering algorithms, decision tree algorithms, regularization algorithms, regression algorithms, example-based algorithms, artificial neural network algorithms, deep learning algorithms, dimensionality reduction algorithms, etc. Various examples of specific algorithms include Bayesian linear regression, enhanced decision tree regression and neural network regression, back propagation neural networks, self-attention models, convolutional neural networks (CNNs), long short-term networks (LSTMs), prior algorithms, K-means clustering, K-nearest neighbors (kNNs), learning vector quantization (LVQ), self-organizing maps (SOMs), local weighted learning (LWL), ridge regression, least absolute shrinkage and selection operator (LASSO), elastic net and least angle regression (LARS), principal component analysis (PCA), and principal component regression (PCR).

[0111] In some examples, the model configuration service 332 may be configured to train the machine learning model 354 using the training data 355. By training the machine learning model based on the training data 355, the model configuration service 332 is configured to cause the machine learning model 354 to identify one or more patterns corresponding to a measure of LV dysfunction (e.g., LVEF) and one or more characteristics of the cardiac electrical data. For example, the model configuration service 332 is configured to cause the machine learning model 354 to identify one or more patterns of one or more values ​​of the measure of LV dysfunction in the cardiac electrical data. This may allow the HMS 26 to use the machine learning model 354 to process the incoming cardiac electrical training data samples to determine the value of the measure of LV dysfunction.

[0112] Training data 355 may include cardiac electrical training data 356 and LV dysfunction metric training data 358. In some examples, training data 355 may include a plurality of training data sets, wherein each training data set in the plurality of training data sets includes a cardiac electrical training data set of cardiac electrical training data 356 and an LV dysfunction metric training data set of LV dysfunction metric training data 358. In some examples, the cardiac electrical training data set and the LV dysfunction metric training data set of the training data sets may correspond to the same patient. That is, the cardiac electrical training data set may be collected from a patient, and the LV dysfunction metric training data set may include one or more values ​​of a metric of LV dysfunction measured from the patient.

[0113] To train the machine learning model 354, the model configuration service 332 may label and / or classify each of the plurality of training data sets of the training data 355. In some examples, to label each of the training data sets of the training data 355, the model configuration service 332 may label the cardiac electrical training data set of each of the training data sets with one or more characteristics, reference points, and / or features of the cardiac electrical training data set. In some examples, to label each of the training data sets of the training data 355, the machine learning model 354 may label the cardiac electrical training data set of each of the training data sets with one or more parameters corresponding to the cardiac electrical training data set, such as heart rate, heart rate variability, pulse transit time, or any combination thereof. In some examples, to label each of the training data sets of the training data 355, the machine learning model 354 may label the cardiac electrical training data set of each of the training data sets with one or more arrhythmias or patient conditions corresponding to the cardiac electrical training data set, such as HF, AV, AF, AV block, or any combination thereof. In some examples, data analyzer 330 may identify one or more characteristics, reference points, features, parameters, patient conditions, arrhythmias, or any combination thereof corresponding to each set of cardiac electrical training data 356 .

[0114] In some examples, to label each training data set of training data 355, model configuration service 332 may identify one or more values ​​of a measure of LV dysfunction in an LV dysfunction metric training data set of LV dysfunction metric training data 358 corresponding to each training data set of training data 355. In some examples, each of the one or more values ​​of the measure of LV dysfunction may correspond to a time or time range at which a value of the measure of LV dysfunction was measured from a patient. Model configuration service 332 may associate a time of each of the one or more values ​​of the measure of LV dysfunction corresponding to the LV dysfunction metric training data set with a time or time range of a corresponding cardiac electrical training data set. For example, since each training data set of training data 355 includes a cardiac electrical training data set of cardiac electrical training data 356 and an LV dysfunction metric training data set of LV dysfunction metric training data 358 collected from the same patient, model configuration service 332 may associate a time at which the cardiac electrical training data was collected from the patient with a time at which each of the one or more values ​​of the measure of LV dysfunction was measured from the patient. This may allow model configuration service 332 to identify a pattern of values ​​of the measure of LV dysfunction in the cardiac electrical training data that indicates a pattern of values ​​of the measure of LV dysfunction.

[0115] In some examples, the model configuration service 332 may create multiple training data sets 355, each of which includes a cardiac electrical training data set (e.g., ECG data and / or EGM data) and an LV dysfunction metric training data set. In some examples, the cardiac electrical training data set and the LV dysfunction metric training data set may be collected from the same patient at different times. For example, the cardiac electrical training data set and the LV dysfunction metric training data set may be collected from the same patient less than a day apart, less than a week apart, less than two weeks apart, or another amount of time apart. In some examples, the cardiac electrical training data set and the LV dysfunction metric training data set may be collected from the same patient at the same time and / or during overlapping time windows. In any case, the model configuration service 332 may, for each training data set of training data 355, associate a time or time range of a corresponding cardiac electrical training data set corresponding to the cardiac electrical training data 356 with a time or time range of a corresponding LV dysfunction metric training data set corresponding to the LV dysfunction metric training data 358. This may allow the model configuration service 332 to train the machine learning model 354 to recognize patterns in the values ​​of the metric indicative of LV dysfunction in the cardiac electrical data.

[0116] In some examples, the cardiac electrical training data 356 may include multiple cardiac electrical training data sets. The multiple cardiac electrical data sets may include one or more cardiac electrical measurements each collected from a human patient via a Holter monitor, one or more cardiac electrical data sets collected via electrodes attached to the skin of a human patient, one or more cardiac electrical data sets collected from a patient via a wearable device (e.g., a smartwatch), one or more cardiac electrical data sets collected from a human patient via an IMD, or any combination thereof. In any case, each of the multiple cardiac electrical training data sets within the cardiac electrical training data 356 may include electrical data indicative of cardiac activity of the patient, such as one or more cardiac cycles of the patient's myocardium.

[0117] In some examples, model configuration service 332 may train machine learning model 354 based on one or more raw cardiac electrical training data sets of cardiac electrical training data 356. The raw cardiac electrical training data may indicate one or more features, characteristics, and / or parameters of cardiac activity without including labels identifying those features or characteristics. For example, the raw cardiac electrical training data set may indicate one or more R waves, T waves, P waves, and other features or characteristics without including labels identifying those features that the raw cardiac electrical training data set may indicate. Additionally or alternatively, the raw cardiac electrical training data set may indicate parameters such as heart rate and / or heart rate variability without including labels identifying the values ​​of those parameters.

[0118] In some examples, the model configuration service 332 may train the machine learning model 354 based on one or more labeled raw cardiac electrical training data sets of cardiac electrical training data 356. The labeled cardiac electrical training data may include labels that identify one or more features, characteristics, and / or parameters of the cardiac activity indicated by the cardiac electrical training data. For example, the labeled cardiac electrical training data set may include one or more labels that identify one or more R waves, T waves, P waves, and other features or characteristics. Additionally or alternatively, the labeled cardiac electrical training data set may include one or more labels that identify parameter values ​​corresponding to the labeled cardiac electrical training data, such as heart rate and / or heart rate variability.

[0119] In some examples, to train the machine learning model 354, it may be beneficial for the model configuration service to use information corresponding to the training data in order to associate patterns and aspects of the training data with known characteristics of the training data. For example, when the model configuration service 332 is training the machine learning model 354 to identify fiducials in cardiac signals (such as ECG signals and EGM signals), it may be beneficial for the training data to include multiple cardiac data samples with labels that identify the fiducials (e.g., labels that identify P waves, R waves, T waves, and other fiducials). Additionally or alternatively, when the model configuration service 332 is training the machine learning model 354 to identify arrhythmias (such as AF or AV block) in the cardiac data, it may be beneficial for the training data to include one or more labels that indicate portions of the training data that indicate arrhythmias.

[0120] In some examples, model configuration service 332 may train machine learning model 354 to process cardiac electrical data to determine a value for LVEF. In some examples, LVEF may represent a ratio of the volume of blood expelled from the left ventricle in response to ventricular depolarization to the volume of blood present in the left ventricle immediately before ventricular depolarization. LVEF may be expressed as a percentage (e.g., 50% of the blood present in the left ventricle before depolarization is expelled from the ventricle in response to depolarization). In some examples, when machine learning model 354 outputs a percentage and / or ratio value for LVEF, machine learning model 354 may represent a regression model.

[0121] In some examples, when the machine learning model 354 outputs a confidence that the patient's LVEF is low (e.g., <35%), the model configuration service 332 may represent a classification model. For example, the machine learning model 354 may process the cardiac electrical data 352 and output a confidence that the value of the measure of LV dysfunction (e.g., the value of LVEF) is less than a threshold value of the measure of LV dysfunction. In some examples, the confidence output by the machine learning model 354 may represent a probability in a range between 0 and 1, i.e., a probability corresponding to the patient's cardiac electrical data 352 indicating that the patient is associated with a value of the measure of LV dysfunction that is below the threshold value of the measure of LV dysfunction. In some examples, the machine learning model 354 may process the cardiac electrical data 352 and output a confidence that the value of the measure of LV dysfunction (e.g., the value of LVEF) is not less than the threshold value of the measure of LV dysfunction.

[0122] The machine learning model 354 may be configured to process cardiac electrical data from any patient's cardiac electrical data 352 to determine a value for a measure of LV dysfunction. For example, the machine learning model 354 may process a cardiac electrical data set corresponding to a patient at rest to determine a value for a measure of LV dysfunction. In other examples, the machine learning model 354 may process a cardiac electrical data set corresponding to an active patient to determine a value for a measure of LV dysfunction. In other examples, the machine learning model 354 may process a cardiac electrical data set corresponding to a patient exhibiting signs of heart failure to determine a value for a measure of LV dysfunction. The machine learning model 354 may process a cardiac electrical data set corresponding to a patient under any conditions to determine a value for a measure of LV dysfunction.

[0123] In some examples, model configuration service 332 may use paired cardiac electrical training data 356 and LV dysfunction metric training data 358 (e.g., time point ECG-LVEF paired data) to develop machine learning model 354 as a classification model to determine whether the value of the metric of LV dysfunction is low or high (e.g., LVEF < 35%). In some examples, model configuration service 332 may use paired cardiac electrical training data 356 and LV dysfunction metric training data 358 (e.g., time point ECG-LVEF paired data) to develop machine learning model 354 as a regression model to estimate the precise value of the metric of LV dysfunction (e.g., LVEF = 64%). When machine learning model 354 is a classification model, machine learning model 354 may output a confidence that the value of the metric of LV dysfunction is below and / or above a threshold. When machine learning model 354 is a regression model, machine learning model 354 may output a precise estimate of the value of the metric of LV dysfunction.

[0124] In some cases, the model configuration service 332 may collect longitudinal cardiac electrical training data (e.g., longitudinal ECG data), as well as time point pairs of cardiac electrical data and LV dysfunction metric data, and longitudinal LV dysfunction metric data (e.g., longitudinal EF data) to develop the machine learning model 354 as a classification model and / or a regression model. In one example, the machine learning model 354 may be a time t 1 ,t 2 …t n ECG data set of cardiac electrical training data 356 (e.g., ECG (t 1 )、ECG(t 2 )…ECG(t n )) vs. time t m LVEF dataset of LV dysfunction metric training data 358 (e.g., LVEF(t m)) to create a training data set of training data 355. The paired sets of cardiac electrical training data and LV dysfunction metric training data may correspond to different time points, the same time points, overlapping time points, and / or open time windows, or any combination thereof.

[0125] In some examples, model configuration service 332 may train machine learning model 354 as a classification model configured to output a confidence that LVEF is below a threshold value of LVEF. In some examples, the threshold value of LVEF is 35%, but this is not required. The threshold value of LVEF may be any value. In some examples, when the LVEF is below the threshold value of LVEF, this may indicate that the patient is at risk of experiencing one or more conditions, such as heart failure. In some examples, the LVEF of a patient may vary depending on one or more factors, such as activity level, posture, whether the patient is awake or asleep, and other factors.

[0126] When the model configuration service 332 trains the machine learning model 354 as a classification model, the model configuration service 332 may use a time point pairing set of cardiac electrical training data and LV dysfunction metric training data. That is, the model configuration service 332 may train the machine learning model 354 using a plurality of training data sets of training data 355, each of which includes a cardiac electrical training data set of cardiac electrical training data 356 and a LV dysfunction metric training data set of LV dysfunction metric training data 358 collected from a patient at the same time point, during the same time window, or during an overlapping time window. When the model configuration service 332 trains the machine learning model 354 as a classification model configured to output a confidence level that the LVEF is below a threshold value of the LVEF using the time point pairing training data, the machine learning model 354 may process the cardiac electrical data of the cardiac electrical data 352 collected from the patient and output a confidence level that the patient's LVEF is below a threshold value of the LVEF at the time or time window when the cardiac electrical data was collected from the patient.

[0127] In some examples, the machine learning model 354 may accept as input an ECG data set of cardiac electrical data 352 corresponding to time k and output an estimate of LVEF at time m. The machine learning model 354 may operate according to the following equation:

[0128] EF(t m ):F(ECG(t k ))~=EF(t m ), where (mk) < 1 week (Equation 1)

[0129] As seen above in Equation 1, the machine learning model 354 may accept as input an ECG dataset of cardiac electrical data 352 corresponding to time k, and output an estimate of the LVEF at time m, where time k occurs less than one week before time m. That is, the model configuration service 332 may train the machine learning model 354 based on paired sets of cardiac electrical training data 356 and LV dysfunction metric training data 358. The cardiac electrical dataset corresponding to each paired set of cardiac electrical training data and LV dysfunction metric training data may be collected from the patient less than one week before the LV dysfunction metric training dataset was collected from the patient. This means that once the machine learning model 354 is trained, the machine learning model 354 may accept as input an ECG dataset of cardiac electrical data 352 collected from the patient, and output a confidence level that a value of the metric of LV dysfunction for the patient will be less than a threshold value for the metric of LV dysfunction less than one week after the ECG dataset was collected from the patient.

[0130] In some examples, the model configuration service 332 may train the machine learning model 354 based on one or more training data sets, each of which includes two or more ECG data sets of cardiac electrical training data 356 collected from a patient paired with an LV dysfunction metric training data set of LV dysfunction metric training data 358 collected from the patient. In some examples, each of the two or more ECG data sets may correspond to a time point or a time window, and the LV dysfunction metric training data set may correspond to a time point or a time window. In some examples, the two or more ECG data sets may each be collected from the patient before the LV dysfunction metric training data set is collected from the patient. In some examples, the two or more ECG data sets may each be collected from the patient after the LV dysfunction metric training data set is collected from the patient. In some examples, some of the two or more ECG data sets may be collected from the patient before the LV dysfunction metric training data set is collected from the patient, and some of the two or more sets of ECG data may be collected from the patient after the LV dysfunction metric training data set is collected from the patient.

[0131] This means that once the machine learning model 354 is trained, the machine learning model 354 may output a value for a measure of LV dysfunction based on two or more ECG datasets of cardiac electrical data 352 collected from a patient. In some examples, when the machine learning model 354 is trained based on a training dataset that includes more than one ECG dataset collected from the same patient, the machine learning model 354 may be more robust than when the machine learning model 354 is trained based on a training dataset that includes a single ECG dataset. In some examples, the machine learning model 354 may operate according to the following equation:

[0132] F(ECG(t k ),ECG(t k -1),ECG(t k -2),ECG(t k +1))~EF(tm)

[0133] As seen above in Equation 2, the machine learning model 354 (F) may accept as input an ECG dataset of cardiac electrical data 352 collected from a patient at time k, an ECG dataset of cardiac electrical data 352 collected from a patient at time k-1, an ECG dataset of cardiac electrical data 352 collected from a patient at time k-2, and an ECG dataset of cardiac electrical data 352 collected from a patient at time k+1. The machine learning model 354 may output a value corresponding to a measure of LV dysfunction for the patient at time m. In some examples, the value of the measure of LV dysfunction may represent a classification of whether the measure of LV dysfunction is less than a threshold measure of LV dysfunction. In some examples, the value of the measure of LV dysfunction may represent an accurate estimate of the value of the measure of LV dysfunction. In some examples, time k, time k-1, time k-2, and time k+1 may each be less than a week before time m.

[0134] In some examples, the model configuration service 332 may train the machine learning model 354 using longitudinal data to output an ECG-based classification of whether the LVEF is low or high (e.g., a confidence level that the LVEF is below a LVEF threshold). In some examples, the model configuration service 332 may train the machine learning model 354 using longitudinal data to output an ECG-based determination of whether the LVEF has increased significantly or decreased significantly. For example, the machine learning model 354 may receive one or more ECG data sets collected from a patient and output a determination of whether the patient's LVEF has increased by more than an LVEF increase threshold or decreased by more than an LVEF decrease threshold.

[0135] In some examples, data analyzer 330 may identify one or more parameters and / or characteristics indicated by cardiac electrical training data 356, such as heart rate, heart rate variability, arrhythmia, R wave, T wave, P wave, or any combination thereof. Model configuration service 332 may develop machine learning model 354 to perform one or more LVEF classifications and / or one or more EF trend detections based on parameter ranges. For example, model configuration service 332 may train machine learning model 354 to process input cardiac electrical data differently based on heart rate, heart rate variability, or arrhythmia.

[0136] In some examples, the LVEF of a patient does not change instantaneously over a short period of time, and changes in the LVEF may not occur frequently. This means that there may be a hysteresis associated with the increasing and decreasing trends of the LVEF. The data analyzer 330 can quantify two parameters from the longitudinal EF data, the hysteresis during EF decrease and the hysteresis during EF increase. These two parameters can be used as parameters for ECG-based EF change detection to detect physiologically true EF changes while reducing the frequency of false positives.

[0137] In some examples, the ICM ECG-LVEF paired measurements of the training data 355 may be limited, but the paired Holter monitor ECG-LVEF measurements of the training data 355 may be more readily available. In some examples, the training data 355 may include one or more paired Holter monitor / wearable ECG-LVEF data sets. In some examples, the training data 355 may include one or more paired Holter monitor / wearable ECG-ICM ECG data sets. That is, one or more training data sets 355 may include ECGs collected from patients via a Holter monitor, a wearable device, an IMD (e.g., an ICM), or any combination thereof. For example, the HMD 26 may map M from the wearable / Holter monitor ECG data to the ICM ECG. In some examples, the HMD 26 may map the ICM ECG to the Holter monitor ECG according to the equation ICM ECG=M(Holter monitor ECG). In some examples, HMD 26 may map the ICM ECG to the Holter monitor ECG according to the equation ICM ECG = 0.75 * Lead II Holter monitor ECG + 0.25 * Lead I Holter monitor ECG. This mapping may be used for dataset D1 to train machine learning model 354 to map M(D1 Holter monitor ECG) ~ = D1 EF. After machine learning model 354 is trained, machine learning model 354 may be used for subsequent ICM ECG-based EF classification.

[0138] The mortality rate of patients with heart failure may be related to the contractile reserve. The term "contractile reserve" may refer to the difference between myocardial contractility and myocardial pressure. In some examples, a higher contractile reserve may be associated with a lower mortality rate, and a lower contractile reserve may be associated with a higher mortality rate. HMS26 may estimate the contractile reserve of a patient with heart failure based on ECG-derived LVEF, ECG-derived autonomic nervous tension, drug information, activity information, or any combination thereof. In other words, HMS26 may monitor a measure of LV dysfunction (e.g., LVEF) of patient 4 based on cardiac electrical data 352 (e.g., ECG data) to determine the contractile reserve of patient 4, thereby monitoring the risk of death of patient 4.

[0139] The LVEF derived from the ECG may represent a measure for monitoring HF deterioration in patients with a reduced LVEF. For example, when patient 4 has a reduced LVEF, HMS26 may track the LVEF of patient 4 based on cardiac electrical data 352 collected from patient 4. HMS26 may identify one or more trends in LVEF based on cardiac electrical data 352 collected from patient 4. When one or more trends in LVEF indicate that the LVEF is reduced, this may indicate that heart failure is deteriorating. When one or more trends in LVEF indicate that the LVEF is increased, this may indicate that heart failure is improving. In some examples, the medical device system 2 may use a multi-lead ICM system to collect cardiac electrical data. In order to maintain battery life, ICM may be mainly used in a single-channel arrhythmia detection mode. For LVEF detection, a multi-channel ECG may be routinely recorded (e.g., once a day, once a week, or according to any other interval) via post-processing for LVEF classification.

[0140] A system for ECG-based EF monitoring may include a single-lead ICM device. During the implantation procedure, a real-time LVEF estimation metric may be calculated to determine whether the LVEF can be estimated based on data collected by the ICM in the implantation orientation. Since arrhythmia monitoring can work across multiple orientations of the ICM, the implantation process can be optimized for ECG-based EF determination. In some examples, the system for ECG-based LVEF monitoring may not analyze conventional ECG events from a device for LVEF estimation on the cloud (e.g., on the computing system 20), but the system may use a sensitive onboard algorithm (e.g., onboard IMD10) to detect low LVEF or detect significant LVEF changes. The ECG collected by the IMD 10 may be post-processed by the HMD 26 to reduce false positives.

[0141] HMS26 can use ICM-ECG-based low LVEF detection to identify one or more patients who need further diagnosis and monitoring with echocardiography and / or cardiac MRI. A system for ECG-based LVEF may include a configurable device in terms of the number or frequency of ECG measurements (e.g., on demand or once per time period), the number of leads, and the ECG sampling rate. This allows the system to switch EF monitoring modes according to patient monitoring needs (e.g., low-frequency monitoring before hospitalization for HF prediction, high-frequency monitoring during hospitalization, and high-frequency monitoring 1 week after discharge). HMS26 can identify whether a HF patient has a low or high LVEF for appropriate treatment follow-up.

[0142] Figure 6is a conceptual diagram illustrating an example machine learning model 400 configured to output information corresponding to cardiac electrical data in accordance with one or more techniques of this disclosure. Machine learning model 400 is an example of a deep learning model or deep learning algorithm. One or more of IMD 10, computing device 12, or computing system 20 (e.g., model configuration service 272 and / or model configuration service 332) may train, store, and / or utilize machine learning model 400, although in other examples, other devices may apply inputs associated with a particular patient to machine learning model 400. Some non-limiting examples of machine learning techniques include Bayesian probability models, Hawkes processes, support vector machines, K-nearest neighbor algorithms, and multilayer perceptrons.

[0143] like Figure 6 As shown in the example of , the machine learning model 400 may include an input layer 402, a hidden layer 404, and an output layer 406. The output layer 406 includes the output of the transfer function 405 from the output layer 406. The input layer 402 represents each of the input values ​​X1 to X4 provided to the machine learning model 400. The number of inputs may be less than or greater than 4, including much greater than 4, such as hundreds or thousands. In some examples, the input values ​​may be parameters determined based on the cardiac electrical data 184, 292, 352 (including those described herein) and in some cases other parameter data 182, 290, 350.

[0144] Each of the input values ​​of each node in the input layer 402 is provided to each node in the hidden layer 404. Figure 6 In the example of , hidden layer 404 includes two layers, one layer having four nodes and the other layer having three nodes, but fewer or greater numbers of nodes may be used in other examples. Each input from input layer 402 is multiplied by a weight and then added at each node of hidden layer 404. During training of machine learning model 400, the weight of each input is adjusted to establish a relationship between the input physiological parameter value and one or more output values ​​indicative of the patient's health state. In some examples, one hidden layer may be incorporated into machine learning model 400, or three or more hidden layers may be incorporated into machine learning model 400, where each layer includes the same or different number of nodes.

[0145] The results for each node within hidden layer 404 are applied to a transfer function of output layer 406. The transfer function may be linear or nonlinear, depending on the number of layers within machine learning model 400. Example nonlinear transfer functions may be a sigmoid function or a rectifier function. The output 407 of the transfer function may be one or more values ​​indicating a classification of whether the measure of LV dysfunction is less than a threshold and / or a determination of the exact value of the measure of LV dysfunction. By applying the data to a machine learning model such as machine learning model 400, the processing circuitry of medical device system 2 is able to determine information corresponding to patient 4 with high specificity and sensitivity.

[0146] Figure 7 5 is a block diagram illustrating an example of a machine learning model 400 trained using supervised learning and / or reinforcement learning in accordance with one or more techniques of the present disclosure. The machine learning model 400 may be implemented using any number of models for supervised learning and / or reinforcement learning, such as, but not limited to, an artificial neural network, a decision tree, a naive Bayesian network, a support vector machine, or a k-nearest neighbor model, to name a few examples. In some examples, processing circuitry of one or more of IMD 10, external device 12, and / or computing system 20 initially trains the machine learning model 400 based on training data 500. In some examples, the training data may include Figure 5 The output of the machine learning model 400 may be compared 504 to a target output 503 determined, for example, based on the labels. Based on an error signal representing the comparison, processing circuitry implementing the learning / training functionality 505 may send or apply modifications to the weights of the machine learning model 400, or otherwise modify / update the machine learning model 400. For example, one or more of the IMD 10, external device 12, and / or computing system 20 may modify the machine learning model 400 in response to the data applied to the machine learning model 400 to change the output generated by the machine learning model 400 for each training instance in the training set 500.

[0147] Figure 8 is a flow chart illustrating an example method for determining whether to use a machine learning model to process cardiac electrical data collected from a patient in accordance with one or more techniques of this disclosure. Figure 8 Relative to Figure 1 The medical device system 2 is described. However, Figure 8 The techniques may be performed by different components of medical device system 2 or by additional or alternative medical device systems.

[0148] HMS26 can train models and personalize and / or calibrate models for specific patients. For example, HMS26 can train machine learning models to implement ECG-based algorithms (e.g., for low LVEF classification) and verify machine learning models with large data sets. HMS26 can apply machine learning models to cardiac electrical data collected from patients for expected use. In some examples, HMS26 can obtain the patient's ECG and paired EF. When the machine learning model is fully performed relative to the patient, the machine learning model can be used according to its current state. When the machine learning model is not fully performed relative to the patient, the machine learning model can be adjusted using ECG-LVEF paired data measurements from the patient and up to more than N ECG-LVEF paired data measurements. When patient-specific data can be used to adjust algorithm parameters for the patient to achieve sufficient performance, a personalized algorithm can be used for the patient. If the machine learning model has been updated for N iterations and is still insufficient, the machine learning model can be marked as not suitable for the patient, and the data can be sorted for overall model updates. "Sufficient performance" can correspond to the ability to detect low LVEF or the ability to detect significant changes in LVEF.

[0149] HMS26 may receive training data including cardiac electrical data and LV dysfunction metric data (802). In some examples, the cardiac electrical data and LV dysfunction metric data may correspond to a patient, such as Patient 4. HMS26 may train a machine learning model based on the cardiac electrical data and LV dysfunction metric data corresponding to Patient 4 (804). In some examples, HMS26 may train the machine learning model based on a training data set that includes one or more paired sets of cardiac electrical data and LV dysfunction metric data corresponding to one or more human patients other than Patient 4. In some examples, HMS26 may receive an incoming cardiac electrical data set from Patient 4.

[0150] HMS 26 may apply a machine learning model to incoming cardiac electrical data corresponding to patient 4 (806) to determine a value of a measure of LV dysfunction. In some examples, the machine learning model may represent a classification model configured to output a confidence that the value of the measure of LV dysfunction is below a threshold for the measure of LV dysfunction. In some examples, the machine learning model may represent a regression model configured to output an estimate of the exact value of the measure of LV dysfunction.

[0151] HMS 26 may determine, based on the output from the machine learning model, whether the machine learning model is adequately performing with respect to patient 4 (808). In some examples, to determine whether the machine learning model is adequately performing with respect to patient 4, HMS 26 may compare the output from the machine learning model with LV dysfunction metric data corresponding to patient 4. When the machine learning model is adequately performing with respect to patient 4 ("yes" at block 808), HMS 26 may process cardiac electrical data corresponding to patient 4 using the machine learning model.

[0152] When the machine learning model does not adequately perform with respect to patient 4 ("No" at box 808), HMS26 may determine whether the machine learning model has been updated at least N times (810). When the machine learning model has been updated at least N times ("Yes" at box 810), HMS26 may refuse to use the machine learning model to process data corresponding to patient 4 (812). When the machine learning model has not been updated at least N times ("No" at box 810), HMS26 may update the machine learning model (814). In some examples, to update the machine learning model, HMS26 may retrain the machine learning model using additional training data corresponding to patient 4. In some examples, to update the machine learning model, HMS26 may retrain the machine learning model using additional training data corresponding to one or more patients other than patient 4. When HMS26 updates the machine learning model, Figure 8 The process may return to block 808.

[0153] Fig. 9 is a flow chart illustrating an example method for applying a machine learning model to determine a value for a measure of LV dysfunction in accordance with one or more techniques of the present disclosure. Fig. 9 Relative to Figure 1 The medical device system 2 is described. However, Fig. 9 The techniques may be performed by different components of medical device system 2 or by additional or alternative medical device systems.

[0154] IMD 10 may generate cardiac electrical data based on cardiac signals sensed by IMD 10 via one or more electrodes (902). HMS 26 may apply a machine learning model to the cardiac electrical data to determine a value of a measure of LV dysfunction (904). The technology disclosed herein is not limited to HMS 26 that applies a machine learning model to cardiac electrical data. In some cases, computing devices 12A, 12B and / or IMD 10 may apply a machine learning model. In some examples, the machine learning model is trained based on multiple training data sets. Each of the multiple training data sets may include a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set. HMS 26 may output the determined value of the measure of LV dysfunction to a computing device (e.g., computing devices 12A, 12B) (906).

[0155] The following numbered clauses may demonstrate one or more aspects of the present disclosure.

[0156] Clause 1: A medical device system includes a medical device system, the medical device system including: a medical device, the medical device including one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes; a memory, the memory configured to store a machine learning model and a plurality of training data sets; and a processing circuit, the processing circuit communicating with the memory. The processing circuit is configured to apply the machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction. The machine learning model is trained based on the plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicating one or more values ​​of a measure of LV dysfunction corresponding to the training cardiac electrical data set. In addition, the processing circuit is configured to output the determined value of the measure of LV dysfunction to a computing device.

[0157] Item 2: A medical device system according to Item 1, wherein the cardiac electrical data includes multiple cardiac electrical data sets, and wherein in order to apply the machine learning model to the cardiac electrical data to determine the value of the measure of LV dysfunction, the processing circuit is configured to: apply the machine learning model to a cardiac electrical data set in the multiple cardiac electrical data sets to determine the value of the measure of LV dysfunction corresponding to the cardiac electrical data set in the multiple cardiac electrical data sets, and wherein the processing circuit is further configured to: apply the machine learning model to each other cardiac electrical data set in the multiple cardiac electrical data sets to determine the value of the measure of LV dysfunction corresponding to each other cardiac electrical data set in the multiple cardiac electrical data sets; and output the determined value of the measure of LV dysfunction corresponding to each other cardiac electrical data set in the multiple cardiac electrical data sets to the computing device.

[0158] Clause 3: A medical device system according to any one of clauses 1 to 2, wherein the medical device further includes an accelerometer, wherein the medical device is further configured to generate motion data based on a motion signal sensed by the accelerometer, and wherein the processing circuit is further configured to: determine a motion value indicating the patient's activity level based on the motion data, wherein the motion value corresponds to a determined value of the measure of the LV dysfunction; and determine whether to output a warning based on the motion value and the determined value of the measure of the LV dysfunction.

[0159] Item 4: A medical device system according to Item 3, wherein, in order to determine whether to output the warning, the processing circuit is configured to: determine whether the value of the measure of LV dysfunction is lower than a threshold measure of LV dysfunction; determine whether the motion value is greater than a threshold motion value; and determine whether to output the warning based on whether the value of the measure of LV dysfunction is lower than the threshold measure of LV dysfunction and whether the motion value is greater than the threshold motion value.

[0160] Clause 5: A medical device system according to any one of clauses 1 to 4, wherein the processing circuit is further configured to train the machine learning model based on the multiple training data sets, and wherein by training the machine learning model based on the multiple training data sets, the processing circuit is configured to cause the machine learning model to identify one or more patterns corresponding to the measure of the LV dysfunction and one or more characteristics of the cardiac electrical data.

[0161] Clause 6: The medical device system of any of clauses 1 to 5, wherein the processing circuit is further configured to label each training data set of the plurality of training data sets.

[0162] Item 7: A medical device system according to Item 6, wherein, in order to label each of the multiple training data sets, the processing circuit is configured to: for each of the multiple training data sets, identify one or more characteristics of the training cardiac electrical data set in the training data set; and label the training cardiac electrical data set of each of the multiple training data sets using the one or more characteristics of the training cardiac electrical data set.

[0163] Clause 8: A medical device system according to Clause 7, wherein the one or more characteristics of the training cardiac electrical data set include any one or more of the following: one or more R waves, one or more P waves, one or more T waves, a heart rate corresponding to the training cardiac electrical data set, a heart rate variability corresponding to the training cardiac electrical data set, and an arrhythmia indicated by the training cardiac electrical data set.

[0164] Clause 9: A medical device system according to any one of clauses 6 to 7, wherein in order to label each of the multiple training data sets, the processing circuit is configured to: identify, in the information indicating one or more values ​​of the measure of LV dysfunction of the training data set, a time corresponding to each of the one or more values ​​of the measure of LV dysfunction; and associate the time corresponding to each of the one or more values ​​of the measure of LV dysfunction with the time of the training cardiac electrical data set.

[0165] Clause 10: A medical device system according to any one of clauses 1 to 9, wherein in order to apply the machine learning model to the cardiac electrical data to determine the value of the LV dysfunction, the processing circuit is configured to apply the machine learning model to the cardiac electrical data to determine the confidence that the value of the LV dysfunction is below a threshold value for LV dysfunction.

[0166] Clause 11: The medical device system of any of clauses 1 to 10, wherein the measure of LV dysfunction comprises ejection fraction.

[0167] Item 12: A method of operating a medical device system comprising a medical device, the medical device comprising one or more electrodes and being configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes, the method comprising: applying a machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction by a processing circuit of the medical device system, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets comprises a training cardiac electrical data set and information indicating one or more values ​​of the measure of the LV dysfunction corresponding to the training cardiac electrical data set, wherein the processing circuit communicates with a memory of the medical device system, the memory being configured to store the machine learning model and the plurality of training data sets; and outputting the determined value of the measure of the LV dysfunction to a computing device by the processing circuit.

[0168] Clause 13: A method according to clause 12, wherein the medical device further includes an accelerometer, wherein the medical device is further configured to generate motion data based on a motion signal sensed by the accelerometer, and wherein the method further includes: determining, by the processing circuit, a motion value indicating the patient's activity level based on the motion data, wherein the motion value corresponds to a determined value of the measure of the LV dysfunction; and determining, by the processing circuit, whether to output a warning based on the motion value and the determined value of the measure of the LV dysfunction.

[0169] Clause 14: A method according to any one of clauses 12 to 13, wherein the medical device further includes an accelerometer, wherein the medical device is further configured to generate motion data based on a motion signal sensed by the accelerometer, and wherein the method further includes: determining, by the processing circuit, a motion value indicating the patient's activity level based on the motion data, wherein the motion value corresponds to a determined value of the measure of the LV dysfunction; and determining, by the processing circuit, whether to output a warning based on the motion value and the determined value of the measure of the LV dysfunction.

[0170] Clause 15: A method according to Clause 14, wherein determining whether to output the warning includes: determining by the processing circuit whether the value of the measure of LV dysfunction is lower than a threshold measure of LV dysfunction; determining by the processing circuit whether the motion value is greater than a threshold motion value; and determining by the processing circuit whether to output the warning based on whether the value of the measure of LV dysfunction is lower than the threshold measure of LV dysfunction and whether the motion value is greater than the threshold motion value.

[0171] Clause 16: A method according to any one of clauses 12 to 15, wherein the method further comprises training the machine learning model based on the multiple training data sets by the processing circuit, and wherein by training the machine learning model based on the multiple training data sets, the method comprises causing the machine learning model to identify, by the processing circuit, one or more patterns corresponding to the measure of the LV dysfunction and one or more characteristics of the cardiac electrical data.

[0172] Clause 17: The method of any one of clauses 12 to 16, wherein the method further comprises labeling, by the processing circuitry, each of the plurality of training data sets.

[0173] Item 18: A method according to Item 17, wherein labeling each of the multiple training data sets comprises: identifying one or more characteristics of the training cardiac electrical data set in the training data set for each of the multiple training data sets by the processing circuit; and labeling the training cardiac electrical data set of each of the multiple training data sets by the processing circuit using the one or more characteristics of the training cardiac electrical data set.

[0174] Clause 19: A method according to any one of clauses 17 to 18, wherein labeling each of the multiple training data sets comprises: identifying, by the processing circuit, in the information indicating one or more values ​​of the measure of LV dysfunction of the training data set, a time corresponding to each of the one or more values ​​of the measure of LV dysfunction; and associating, by the processing circuit, the time corresponding to each of the one or more values ​​of the measure of LV dysfunction with the time of the training cardiac electrical data set.

[0175] Item 20: A non-transitory computer-readable storage medium comprises program instructions which, when executed by a processing circuit of a medical device system comprising one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes, causes the processing circuit to: apply a machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets comprises a training cardiac electrical data set and information indicating one or more values ​​of the measure of the LV dysfunction corresponding to the training cardiac electrical data set, wherein the processing circuit communicates with a memory of the medical device system, the memory being configured to store the machine learning model and the plurality of training data sets; and output the determined value of the measure of the LV dysfunction to a computing device.

[0176] The techniques described in the present disclosure may be implemented at least in part in the form of hardware, software, firmware, or any combination thereof. For example, various aspects of these techniques may be implemented in one or more microprocessors, DSPs, ASICs, FPGAs, or any other equivalent integrated or discrete logic QRS circuits, and any combination of such components, which are embodied in an external device (such as a physician or patient programmer, simulator, or other device). The terms "processor" and "processing circuit" may generally refer to any of the aforementioned logic circuits, alone or in combination with other logic circuits, or any other equivalent circuit, alone or in combination with other digital or analog circuits.

[0177] For various aspects implemented in software, at least some of the functionality attributed to the systems and devices described in this disclosure may be embodied as instructions on a computer-readable storage medium, such as RAM, DRAM, SRAM, magnetic disk, optical disk, flash memory, or various forms of EPROM or EEPROM. The instructions may be executed to support one or more aspects of the functionality described in this disclosure.

[0178] In addition, in some aspects, the functionality described herein can be set within dedicated hardware and / or software modules. Describing different features as modules or units is intended to highlight different functional aspects, and does not necessarily imply that such modules or units must be implemented by separate hardware or software components. On the contrary, the functionality associated with one or more modules or units can be performed by separate hardware or software components, or integrated in common or separate hardware or software components. In addition, these techniques can be fully implemented in one or more circuits or logic elements. The technology disclosed herein can be implemented in various devices or equipment, including an IMD, an external programmer, a combination of an IMD and an external programmer, an integrated circuit (IC) or a group of ICs and / or discrete circuits resident in an IMD and / or an external programmer.

Claims

1. A medical device system, comprising: a medical device comprising one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes; a memory configured to store a machine learning model and a plurality of training data sets; as well as a processing circuit in communication with the memory, wherein the processing circuit is configured to: applying the machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on the plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicative of one or more values ​​of the measure of LV dysfunction corresponding to the training cardiac electrical data set; as well as The determined value of the measure of LV dysfunction is output to a computing device.

2. The medical device system of claim 1 , wherein the cardiac electrical data comprises a plurality of cardiac electrical data sets, wherein to apply the machine learning model to the cardiac electrical data to determine the value of the measure of LV dysfunction, the processing circuit is configured to: applying the machine learning model to a cardiac electrical dataset of the plurality of cardiac electrical datasets to determine the value of the measure of LV dysfunction corresponding to the cardiac electrical dataset of the plurality of cardiac electrical datasets, and Wherein the processing circuit is further configured to: applying the machine learning model to each other of the plurality of cardiac electrical data sets to determine a value of the measure of LV dysfunction corresponding to each other of the plurality of cardiac electrical data sets; and The determined value of the measure of LV dysfunction corresponding to each other cardiac electrical data set of the plurality of cardiac electrical data sets is output to the computing device.

3. The medical device system of claim 1 , wherein the medical device further comprises an accelerometer, wherein the medical device is further configured to generate motion data based on a motion signal sensed by the accelerometer, and wherein the processing circuit is further configured to: determining a motion value indicative of an activity level of the patient based on the motion data, wherein the motion value corresponds to the determined value of the measure of LV dysfunction; and Whether to output a warning is determined based on the motion value and the determined value of the measure of LV dysfunction.

4. The medical device system according to claim 3, wherein in order to determine whether to output the warning, the processing circuit is configured to: determining whether the value of the measure of LV dysfunction is below a threshold measure of LV dysfunction; determining whether the motion value is greater than a threshold motion value; as well as Whether to output the warning is determined based on whether the value of the measure of LV dysfunction is below a threshold measure of LV dysfunction and whether the motion value is greater than the threshold motion value.

5. The medical device system according to claim 1, wherein the processing circuit is further configured to train the machine learning model based on the plurality of training data sets, and Wherein, by training the machine learning model based on the plurality of training data sets, the processing circuit is configured to cause the machine learning model to identify one or more patterns corresponding to the measure of LV dysfunction and one or more characteristics of cardiac electrical data. 6 . The medical device system of claim 1 , wherein the processing circuit is further configured to label each training data set in the plurality of training data sets.

7. The medical device system of claim 6, wherein to label each of the plurality of training data sets, the processing circuit is configured to: for each of the plurality of training data sets, identifying one or more characteristics of the training cardiac electrical data set in the training data set; and The training cardiac electrical data set of each training data set in the plurality of training data sets is labeled using the one or more characteristics of the training cardiac electrical data set.

8. A medical device system according to claim 7, wherein the one or more characteristics of the training cardiac electrical data set include any one or more of the following: one or more R waves, one or more P waves, one or more T waves, a heart rate corresponding to the training cardiac electrical data set, a heart rate variability corresponding to the training cardiac electrical data set, and an arrhythmia indicated by the training cardiac electrical data set.

9. The medical device system of claim 6, wherein to label each of the plurality of training data sets, the processing circuit is configured to: identifying, in the information indicative of the one or more values ​​of the measure of LV dysfunction of the training data set, a time corresponding to each of the one or more values ​​of the measure of LV dysfunction; and The time corresponding to each of the one or more values ​​of the measure of LV dysfunction is associated with a time of the training cardiac electrical data set.

10. A medical device system according to claim 1, wherein in order to apply the machine learning model to the cardiac electrical data to determine the value of the LV dysfunction, the processing circuit is configured to apply the machine learning model to the cardiac electrical data to determine the confidence that the value of the LV dysfunction is below a threshold value of LV dysfunction.

11. The medical device system of claim 1, wherein the measure of LV dysfunction comprises ejection fraction.

12. A method of operating a medical device system comprising a medical device, the medical device comprising one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes, the method comprising: applying, by processing circuitry of the medical device system, a machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicative of one or more values ​​of the measure of LV dysfunction corresponding to the training cardiac electrical data set, wherein the processing circuitry is in communication with a memory of the medical device system, the memory being configured to store the machine learning model and the plurality of training data sets; as well as The determined value of the measure of LV dysfunction is output by the processing circuit to a computing device.

13. The method of claim 12, wherein the medical device further comprises an accelerometer, wherein the medical device is further configured to generate motion data based on motion signals sensed by the accelerometer, and wherein the method further comprises: determining, by the processing circuitry, a motion value indicative of an activity level of the patient based on the motion data, wherein the motion value corresponds to a determined value of the measure of LV dysfunction; as well as Whether to output a warning is determined by the processing circuit based on the motion value and the determined value of the measure of LV dysfunction.

14. The method of claim 12, wherein the medical device further comprises an accelerometer, wherein the medical device is further configured to generate motion data based on motion signals sensed by the accelerometer, and wherein the method further comprises: determining, by the processing circuitry, a motion value indicative of an activity level of the patient based on the motion data, wherein the motion value corresponds to a determined value of the measure of LV dysfunction; as well as Whether to output a warning is determined by the processing circuit based on the motion value and the determined value of the measure of LV dysfunction.

15. The method of claim 14, wherein determining whether to output the warning comprises: determining, by the processing circuitry, whether the value of the measure of LV dysfunction is below a threshold measure of LV dysfunction; determining, by the processing circuit, whether the motion value is greater than a threshold motion value; as well as Whether to output the warning is determined by the processing circuit based on whether the value of the measure of LV dysfunction is below the threshold measure of LV dysfunction and whether the motion value is greater than the threshold motion value.

16. The method according to claim 12, The method further comprises training the machine learning model based on the plurality of training data sets by the processing circuit, and Wherein, by training the machine learning model based on the plurality of training data sets, the method includes causing the machine learning model, through the processing circuit, to identify one or more patterns corresponding to the measure of the LV dysfunction and one or more characteristics of the cardiac electrical data.

17. The method of claim 12, wherein the method further comprises labeling, by the processing circuit, each of the plurality of training data sets.

18. The method of claim 17, wherein labeling each of the plurality of training data sets comprises: identifying, by the processing circuitry, for each of the plurality of training data sets, one or more characteristics of the training cardiac electrical data set in the training data sets; as well as The training cardiac electrical data set of each training data set in the plurality of training data sets is labeled, by the processing circuit, with the one or more characteristics of the training cardiac electrical data set.

19. The method of claim 17, wherein labeling each of the plurality of training data sets comprises: identifying, by the processing circuitry, in the information indicative of the one or more values ​​of the measure of LV dysfunction of the training data set, a time corresponding to each of the one or more values ​​of the measure of LV dysfunction; as well as The time of each of the one or more values ​​of the measure of LV dysfunction is associated, by the processing circuit, with a time of the training cardiac electrical data set.

20. A non-transitory computer-readable storage medium comprising program instructions that, when executed by a processing circuit of a medical device system comprising one or more electrodes and configured to generate cardiac electrical data based on cardiac signals sensed from a patient via the one or more electrodes, cause the processing circuit to: applying a machine learning model to the cardiac electrical data to determine a value of a measure of left ventricular (LV) dysfunction, wherein the machine learning model is trained based on a plurality of training data sets, wherein each of the plurality of training data sets includes a training cardiac electrical data set and information indicative of one or more values ​​of the measure of LV dysfunction corresponding to the training cardiac electrical data set, wherein the processing circuit is in communication with a memory of the medical device system, the memory being configured to store the machine learning model and the plurality of training data sets; and The determined value of the measure of LV dysfunction is output to a computing device.

Images