IT202200026454A1 - AUTOMATIC CARDIAC ARRHYTHMIA DIAGNOSIS SYSTEM - Google Patents

Description

Description of the invention entitled:

?AUTOMATIC DIAGNOSIS SYSTEM FOR CARDIAC ARRHYTHMIAS?

Description

Field of technique

The invention refers to an innovative system for the automatic diagnosis of cardiac arrhythmias from multiple traces, facilitating the monitoring of patients at risk of incurring atrial fibrillation; more specifically, said system allows the extrapolation of the patient's physiological data through a pocket-sized and traceable IoT device.

Known art

The electrocardiogram is a graphic reproduction of the electrical activity of the heart recorded at the surface of the body.

On the surface of the human body (as well as that of any other animal with cardiac muscle), particularly at the trunk level, there are low intensity electric fields which are mainly due to the periodic depolarizations and repolarizations of the heart.

The electrical potentials produced by the cardiac muscle are the sum of minimal potential differences generated by individual cardiac muscle cells. These small voltages can be recorded using a device called an electrocardiograph, modified and improved by Willem Einthoven and ?tienne-Jules Marey in 1903 by direct derivation from a string galvanometer. Many of the rules established by Einthoven remain in modern times and form the basis for interpreting many aspects of the current ECG.

The acquisition of the potential difference by electrodes placed on the body surface occurs thanks to the conductivity of the interstitial fluid of the human body. The acquired signal is translated into an electrocardiographic trace, which represents the easiest, least expensive and most practical method to observe whether the electrical activity of the heart is normal or whether there are pathologies of a mechanical or bioelectrical nature. The normal ECG trace has a characteristic appearance: the trace is characterized by a sequence of positive and negative deflections, called "waves", separated by some straight sections, called "segments". The sequence is repeated at each cardiac cycle.

The normal heartbeat originates thanks to the spontaneous depolarization activity of the cells of the sinoatrial node, a structure located in the right atrium of the heart. The resulting rhythm is defined as sinus, and is characterized by the presence on the surface ECG of particular waves with a regular cadence called P waves. Atrial fibrillation is a cardiac arrhythmia characterized by the absence of P waves on the surface ECG and by a marked variability of the intervals between one heartbeat and the next. The gold standard for the instrumental diagnosis of atrial fibrillation is therefore based on the morphological analysis of the surface ECG, which aims to determine the absence of P waves. In the current state of the art, however, analyzing the morphology of the surface ECG is still cumbersome and complicated. In order to make an accurate diagnosis, it is necessary to install a plurality of of electrodes, the use of a standard electrocardiograph, or surgically implantable recording devices.

In addition, the presence of a doctor who can interpret the results obtained from the ECG trace is essential to identify any anomalies.

Recently, models of the heart in space have become available for viewing and printing 3D models [1].

J.J., 2021. 3D printing in cardiology: A review of applications and roles for advanced cardiac imaging. Annals of 3D Printed Medicine, 4, p.100034. The patients most at risk today are not monitored when they reside at home or when they go out, consequently they are subject to non-negligible risks, such as the onset of atrial fibrillation.

It is necessary to provide a monitoring system that can accurately follow and control the physiological parameters of patients at risk.

Even the utility model CN200942081, published on September 9, 2007, does not solve the above problems, revealing a portable and simplified machine for reading ECG signals, without in any way providing any help in interpreting the patient's results.

The purpose of this patent application is therefore to provide comprehensive assistance that facilitates the self-diagnosis of cardiac arrhythmias, with particular reference to atrial fibrillation, by recording the biosignals on a wearable monitoring device and sending the results obtained to a specialized algorithm, further facilitating the monitoring of patients who require constant checks.

Description of the invention

According to the present invention, an automatic cardiac arrhythmia diagnosis system is created which effectively solves the above-mentioned problems.

The system described in this patent application aims to facilitate the monitoring of patients at risk of atrial fibrillation. It is therefore necessary to constantly and live monitor, remotely, the ECG of these patients, obtaining more information through the use of a pocket-sized IoT device, which can facilitate the interpretation, by doctors, of the physiological data collected.

This self-measurement system is designed to store the collected data in a cloud, allowing for constant improvement of the algorithm to detect the first signs of atrial fibrillation.

The algorithm specialized in identifying atrial fibrillation signals is implemented on the cloud and, through a server, receives as input the biometric data collected by the various sensors available to the patient.

This algorithm has the task of combining a set of morphological parameters extracted from the available biometric data, in order to reveal the occurrence of atrial fibrillation.

The algorithm uses multiple parameters extracted from morphological and temporal biometric data. Furthermore, the algorithm uses an augmented reality representation to combine the information from multiple ECG traces with a spatial volumetric description of the heart in order to represent cardiac activity through spatial-temporal characteristics.

The algorithm uses, in addition to the descriptive data of morphological and temporal information, the following augmented reality information: the information of the spatial model of the heart through a point cloud and the temporal information relating to the ECG traces. Each point (xi, yi, zi), i=1, ...N belonging to the point cloud representing the surface of the heart, is associated with an activation value fi, i=1,..N obtained by adding the radial projection of the multi-lead signals, modified through a nonlinearity, on different portions of the surface of the heart, as shown in Fig. 2 for a single ECG trace.

When the values of ECG signals change with time, the activation function changes with time. Therefore, the information associated with the activation value fi relative to the point (xi, yi, zi), changes with the time index t.

At each time index t, the principal component of rank 1, called u1(t), of the weighted data vector (fi(t) xi, fi(t) yi, fi(t) zi), i=1, ...N, is calculated. This component is calculated according to the L1(L1-PC) norm to obtain a better noise rejection.

The PC L1 vector u1(t) is a vector associated with a spatial direction varying in time. Note that this vector does not coincide with the so-called cardiac vector, being jointly representative of a volumetry and of non-linearly filtered versions of the acquired ECG traces. For the purposes of the subsequent classification, the algorithm calculates the following descriptive temporal trends of u1(t): 1) azimuthal angle ?(t) of u1(t) and 2) polar angle ?(t) of u1(t). These trends are hereinafter referred to as Augmented Reality Coding Information (ARCI).

Biometric data parameters and ARCI trends representing u1(t) are used as characterizing elements of the state of absence or presence of atrial fibrillation and sent to the subsequent machine learning system.

The correct functioning of the system requires a training phase, appropriately managed by the patient under the supervision of a doctor, in which the algorithm is initialized. In this phase, the system is put into operation and the biometric data and the patient's ARCI data are collected over a period of time in which the behavior of the normal cardiac sinus rhythm is observed. These data will then be maintained in the cloud and used as a benchmark to reveal the occurrence of atrial fibrillation when the patient uses the system alone remotely. A plurality of electrodes are then applied directly to the patient's chest and are designed to record the cardiac electrical activity during the daily life of said patient; said electrodes have the purpose of sending the collected ECG signals to the IoT device worn by the patient.

The ECG during the daily life of said patient, by way of example but not limitation, is monitored by a heart rate monitor, suitable for being easily applied to the patient using any elastic band; said heart rate monitor comprising the electrodes for recording the ECG and a Bluetooth device for sending the ECG signals to said IoT device.

Said heart rate monitor, in one embodiment, is designed to include said IoT device inside it.

Said IoT device is worn by the patient thanks to its pocket-sized format and is designed to include a plurality of components inside it, in order to obtain detailed information on the patient's health status, in order to send a detailed report to a server, via wireless connection, thanks to which a specific algorithm can detect the first signs of atrial fibrillation. A GPS receiver, installed inside said IoT device, is designed to monitor the movements of the patient wearing said IoT device and to derive the speed of movement to calibrate the heart rate in relation to the type of physical activity.

A gyroscope, designed to work together with an accelerometer, is further included within the IoT device, in order to determine whether the patient is lying down, standing, or walking or running.

The IoT device further includes a microphone, capable of recording the patient's voice to listen to the tone in order to evaluate their health status.

An EDA sensor is instead installed inside the IoT device to measure the patient's stress levels by detecting small electrical changes on the skin due to electrodermal activity.

All the information collected by the above mentioned sensors is processed by at least one microcontroller, in order to send the detailed report of the physiological data to a server.

Said microcontroller, capable of processing all the information collected by the aforementioned sensors, is constituted, in one embodiment, by an Arduino platform; alternatively, said microcontroller is constituted by a "Raspberry pi Zero" board.

An ESP32 module is used to enable said IoT device to share the information collected by the aforementioned sensors via Bluetooth or Wifi connection.

An ADS1293 module, instead, is suitable for managing the functions typically required in low-power portable ECG applications.

To store the collected ECG signals, before sending them to a server, a micro SD is used, installed internally to the IoT device.

A generic smartphone is used to receive and manage, through a specific application, the detailed report received from said IoT device via Bluetooth technology; said generic smartphone constitutes the bridge between the IoT device and a server. Said central server is able to receive, via an Internet connection, the detailed report containing the ECG signals collected by said electrodes; the server receives the detailed report from the application of said smartphone and is able to send said report to an algorithm that operates in the cloud with the information obtained; said server consequently sends an alarm to the patient if the algorithm detects the presence of atrial fibrillation.

By way of example but not limitation, said server is capable of sending an alarm to any emergency room station if the algorithm detects the presence of atrial fibrillation in a patient monitored by the present system of the invention, further providing the geographical position of the patient thanks to the GPS receiver included within the IoT device.

Furthermore, this server is able to generate an augmented reality representation of cardiac activity for healthcare personnel for diagnostic purposes.

In one embodiment, said server is capable of communicating directly with said IoT device via a low-power LoRaWAN communication network.

The present system of the invention further comprises a cloud, where the detailed reports of each patient are uploaded after being sent to said central server; said cloud comprises all the historically recorded information, forming a database useful for improving the algorithm used for the identification of atrial fibrillation.

During the IoT device's activity period, the electrodes also remain in function, returning a continuous ECG trace over time. It is known that atrial fibrillation is a sporadic cardiac arrhythmia with a limited duration over time. For this reason, the microcontroller takes care of segmenting the received ECG signal into many sub-windows and calculating the projection of the sub-windows onto the volumetric data of the heart represented in the form of a point cloud. These sub-windows can be temporarily stored on the SD card, to then be sent to the cloud and managed by said algorithm.

The advantages offered by the present invention are evident in light of the description set forth thus far and will be even clearer thanks to the attached figures and the detailed description.

Description of the figures

The invention will be described below in at least one preferred embodiment for explanatory and non-limiting purposes with the aid of the attached figures, in which:

- FIGURE 1 shows the generalized architecture of the self-measurement heart rate system of the invention. The monitored patient is in fact equipped with a plurality of electrodes 10 suitable for recording cardiac activity. The ECG signal obtained through said electrodes 10 is sent to an IoT device 18, which, being pocket-sized, is always next to the patient. Said IoT device 18 includes a plurality of sensors including a GPS receiver 12, a gyroscope 13, a microphone 15 and an EDA sensor 17. Said sensors are suitable for detecting the physiological biometric data of the monitored patient, constituting a detailed report on the state of health. Said IoT device 18 includes a micro-SD card 14 and an ESP32 module 16 suitable for allowing said IoT device 18 to share the information collected by said sensors via Bluetooth 22 or Wi-Fi connection. A microcontroller 11 is included in order to manage all the sensors installed internally to said IoT device 18. The detailed report on the patient's health status is subsequently sent, by means of a smartphone 19 which acts as a bridge, to a server 20. Said server 20 in turn sends the detailed report to a cloud 21 where the information is managed and processed by an algorithm specialized in recognizing atrial fibrillation. The server 20 is further able to communicate directly with said IoT device 18 by means of a low-power communication network of the LoRaWAN type 23.

- FIGURE 2 The figure shows the virtual reality representation of the cardiac electrical activity using the spatial model of the heart by means of a point cloud. The electrical activity represented in the generic trace is associated with specific points of the cloud based on pre-established physiological information. - FIGURE 3. The figure shows the representation of the spatial model of the heart by means of a point cloud, the temporal information relating to the ECG traces, and the generation from this information of a vector that geometrically summarises the electrical activity observed over time. This vector is the principal component of rank 1 according to the L1 norm, called u1(t) (L1 PC) of the weighted data vector (fi xi, fi yi, fi zi), i=1, ...N representing the activation level of each point of the cloud. The azimuthal and polar angles of the vector u1(t) constitute the information called Augmented Reality Coding Information.

Detailed description of the invention

The present invention will now be illustrated purely by way of example but not in a limiting or binding manner, with reference to the figure which illustrates some embodiments relating to the present inventive concept.

With reference to FIG. 1, the generic operating scheme of the cardiac arrhythmia self-diagnosis system is illustrated.

A plurality of electrodes 10 must be installed on the chest of the patient to be monitored, in order to extrapolate the electrocardiogram useful to the doctor to determine any cardiac pathologies.

If the patient takes part in any type of physical activity, the system in question is capable of determining the physiological parameters by detecting the position and speed of movement of the patient.

This is possible thanks to a GPS receiver 12 installed inside an IoT device 18, which is designed in a pocket-sized format to always accompany the patient. Said IoT device 18 is designed to include within it the plurality of sensors capable of providing a detailed report on the patient's health status.

In addition to the GPS receiver 12, a gyroscope 13 is installed, which works together with an accelerometer to determine the type of activity the patient is performing, perhaps justifying the accelerated heartbeat.

An EDA sensor 17 is also installed inside the IoT device 18 to measure the patient's stress levels by detecting small electrical changes on the skin due to electrodermal activity.

A microphone 15? is advantageously installed to listen to the patient's tone of voice.

Said IoT device 18 further includes within it a micro-SD card 14 and an ESP32 module 16 capable of allowing said IoT device 18 to share the information collected by the sensors via Bluetooth 22 or Wi-Fi connection.

The collected information composes a detailed report regarding the patient's biometric data, said report is sent, via Bluetooth connection 22, to a generic smartphone 19 which acts as a bridge between said IoT device 18 and a server 20, where the collected reports are intended to be stored.

Said server 20 then sends the patient information to a cloud 21 where an algorithm operates to process the signals and automatically diagnose atrial fibrillation.

The smartphone 19 can also be removed from communications between said IoT device 18 and said server 20, as said components 18, 20, are capable of communicating with each other by means of a low-power communication network of the LoRaWAN type 23.

Finally, it is clear that modifications, additions or variations obvious to a person skilled in the art may be made to the invention described herein, without thereby departing from the scope of protection provided by the attached claims.

Claims (8)

Claims 1. Automatic diagnosis system for cardiac arrhythmias, designed to monitor a patient's cardiac activity and designed for the early diagnosis of atrial fibrillation; said system being designed to send the collected ECG signals to a server (20) that operates remotely; said server (20) using an algorithm dedicated to the diagnosis of atrial fibrillation; all information relating to the patient's health status being sent via wireless connections, created and supported by common devices such as smartphones (19); said server (20) being designed to send the patient an alarm if said algorithm detects the risk of atrial fibrillation; said self-measuring heart rate system being characterised by the fact that it comprises: - a plurality of electrodes (10), directly applicable on the patient's chest, suitable for monitoring the heart rate during the daily life of said patient; said electrodes (10) being suitable for sending the collected ECG signals to an IoT device (18) worn by the patient; - at least one IoT device (18) made in pocket format, wearable by the patient, capable of sending to said server (20), via wireless connection, a detailed report thanks to which a specific algorithm can detect the first signs of atrial fibrillation; - at least one GPS receiver (12), installed internally to said IoT device (18), capable of monitoring the movements of the patient wearing said IoT device (18) and deriving the speed of movement to calibrate the variation of the heartbeat in relation to the type of physical activity; - at least one gyroscope (13), capable of working together with an accelerometer, installed internally to said IoT device (18), in order to determine whether the patient is lying down, standing, or walking or running; - at least one microphone (15), installed internally to said IoT device (18), capable of recording the patient's voice to listen to its tone in order to evaluate its state of health; - at least one EDA sensor (17), installed internally to said IoT device (18), capable of measuring the patient's stress levels, detecting every minimal electrical change on the skin due to electrodermal activity; - at least one microcontroller (11), included internally in said IoT device (18), capable of processing all the information collected by the aforementioned sensors, in order to send a detailed report to said server (20); - at least one ESP32 microcontroller (16) capable of sharing the information collected by the aforementioned sensors via Bluetooth (22) or Wifi connection; - at least one ADS1293 module capable of managing the functions typically required in low-power portable ECG applications; - at least one micro-SD (14), installed internally to said IoT device (18), suitable for storing the collected ECG signals before sending them to said server (20) via wireless connection; - at least one generic smartphone (19), capable of receiving and managing, through a specific application, the detailed report received from said IoT device (18) via Bluetooth technology (22); said generic smartphone (19) acting as a bridge between the IoT device (18) and said server (20) in the event that said IoT device (18) is not capable of sending said information autonomously; - at least one said central server (20), capable of receiving, via wireless connection, the detailed report containing the ECG signals collected by said electrodes (10); the server (20) receiving said detailed report from the application of said smartphone (19) and being capable of sending said report to the algorithm operating in a cloud (21) with the information obtained; said server (20) sending an alarm to the patient if the algorithm detects the risk of atrial fibrillation; - at least one cloud (21), where the detailed reports of each patient are uploaded after being sent to said central server (20); said cloud (21) including all the historically recorded information, forming a database useful for improving the algorithm used for the identification of atrial fibrillation; said algorithm receiving as input the ECG trace measured by the electrodes, which is preprocessed in order to extract from it direct data regarding the measured cardiac rhythm, which are subsequently combined with those obtained from the other sensors (10, 12, 13, 15, 17) to perform the diagnosis of atrial fibrillation. 2. Automatic cardiac arrhythmia diagnosis system, according to the previous claim 1, characterized in that the heart rate during the daily life of said patient is monitored by a heart rate monitor, suitable to be applied to the patient by means of any elastic band; said heart rate monitor comprising a Bluetooth device for sending the ECG signals to said IoT device (18). 3. Automatic cardiac arrhythmia diagnosis system, according to the previous claim 2, characterised in that said heart rate monitor is suitable for including said IoT device (18) inside it. 4. Automatic cardiac arrhythmia diagnosis system, according to any of the preceding claims, characterised in that said microcontroller (11), capable of processing all the information collected by the aforementioned sensors, is made up of a "Raspberry pi Zero" board. 5. Automatic cardiac arrhythmia diagnosis system, according to any of the preceding claims, characterised in that said server (20) is capable of sending an alarm to any emergency room station if the algorithm detects the presence of atrial fibrillation in a monitored patient, further providing the geographical position of the patient thanks to the GPS receiver (12) included within the IoT device (18). 6. Automatic cardiac arrhythmia diagnosis system, according to any of the preceding claims, characterized in that said server (20) is able to communicate directly with said IoT device (18) by means of a low-power communication network of the LoRaWAN type (23). 7. Automatic cardiac arrhythmia diagnosis system, according to any of the preceding claims, characterized in that said server (20) is able to detect fibrillation also by means of the augmented reality representation of the observed ECG traces with a pre-existing 3D cardiac model. 8. Automatic cardiac arrhythmia diagnosis system, according to any of the preceding claims, characterized in that said server (20) is capable of generating an augmented reality view of the observed ECG traces for the purpose of evaluation by healthcare personnel.