Device for processing cardiac electrogram spatiotemporal dispersion signals

Abstract

A device for processing cardiac electrogram spatiotemporal dispersion signals comprises a memory (4) arranged to receive cardiac electrogram spatiotemporal dispersion signals associated, on the one hand, with a time marker and, on the other hand, with a cardiac electrogram trace, a computer (8) arranged to receive, as input, an electrogram trace identifier and a time marker to analyse the cardiac electrogram spatiotemporal dispersion signal associated with this electrogram trace identifier, to analyse a signal extract between the time marker and the first preceding time marker of which the cardiac electrogram spatiotemporal dispersion signal value indicates an absence of dispersion, to obtain from this signal extract, on the one hand, a flatness value of the signal and, on the other hand, a duration value derived from the duration of the signal extract, and to return a trace priority value calculated from the flatness value and the duration value, and a monitor (6) arranged to receive the cardiac electrogram spatiotemporal dispersion signals associated with each cardiac electrogram trace, and, when the value of a cardiac electrogram spatiotemporal dispersion signal indicates a relevant dispersion, to call the computer (8) with the corresponding time marker and the corresponding electrogram trace identifier.

Description

Description Title of the invention: Device for processing spatiotemporal dispersion signals of cardiac electrograms

[0001] The invention relates to the field of signal analysis from cardiac electrograms.

[0002] The treatment of atrial fibrillation has seen significant progress in the last decade. To treat atrial fibrillation, practitioners operate by inserting catheters equipped with multiple electrodes. These electrodes are moved within the heart to measure the electrical signals that propagate through it. The signals obtained are called cardiac electrograms. These cardiac electrograms are processed to help practitioners detect the area or areas of the heart that are causing the atrial fibrillation. Once these areas are identified, the practitioner cauterizes them to render them inactive, thereby restoring normal heart function and eliminating atrial fibrillation.

[0003] Most existing solutions are based on the analysis of CFAEs, or "Complex Fractionated Atrial Electrograms." The principle is to try to find locations in the atria where the electrograms lose their continuity, that is, become fractionated.

[0004] The Applicant has developed an electrogram treatment that has piqued the interest of the scientific community since the publication of very positive clinical studies. These have led to the publication of several articles, including that of Seitz et al. "AF Ablation Guided by Spatiotemporal Electrogram Dispersion Without Pulmonary Vein Isolation: A Wholly Patient-Tailored Approach", Journal of the American College of Cardiology, Volume 69, Issue 3, 24 January 2017, pages 303-321.

[0005] This treatment relies on the detection of a quantity in cardiac electrogram signals called dispersion by the Applicant. More specifically, the Applicant has discovered that measuring the spatiotemporal dispersion of cardiac electrograms—that is, based both on the temporal evolution of the cardiac electrogram signals from each electrode and on the cardiac electrogram signals from neighboring electrodes—is particularly effective in determining the areas of the heart that cause atrial fibrillation. The measurement of this quantity has been granted a patent in numerous countries worldwide and published in Europe under number EP 3 236 843.

[0006] By extending its work, the Applicant incorporated dispersion into a machine learning tool, in order to determine in near real-time the presence of a Dispersion within cardiac electrograms during a procedure allows the practitioner to be alerted, accelerating the ablation procedure and improving the quality of the results obtained, while reducing surgical risks. Furthermore, the reliability of the procedure is increased, limiting the number of patients requiring repeat procedures for the same cause.

[0007] This work has been integrated into software called VX1, which has received FDA (Food and Drug Administration) approval and CE marking. While not providing a diagnosis itself, VX1 assists the practitioner in making a diagnosis: based on the dispersion measurements presented in the software, the practitioner can choose to mark areas of the heart requiring catheter ablation.

[0008] In its current version, the VX1 software uses a database of annotated cardiac electrograms to train a machine learning engine that returns a value indicating the probability that an area for which a cardiac electrogram signal has been analyzed is experiencing atrial fibrillation. Therefore, this is not a measure of dispersion per se, nor information constituting a medical diagnosis, but rather an indication of the greater or lesser chance that a cardiac electrogram reflects the occurrence of atrial fibrillation. Thus, the value returned by the VX1 software is a value between 0 and 1 (0 indicating that there is only a near-zero probability that the area being measured is involved in fibrillation, and 1 indicating that this probability is near certainty).The software is configured so that a value of 0.5 is used as the threshold from which the practitioner's attention is drawn to the potential presence of a dispersion.

[0009] During the development of the VX1 software, the Applicant observed that, in some cases, numerous areas of the heart returned a value greater than 0.5, complicating the practitioner's work, as they must perform as few ablations as possible. Furthermore, when too many areas of dispersion are identified, the practitioner tends to lose confidence in the software's measurement accuracy, since it returns highly noisy information.

[0010] The invention improves the situation. To this end, it proposes a device for processing spatiotemporal dispersion signals of cardiac electrograms, comprising a memory arranged to receive spatiotemporal dispersion signals of cardiac electrograms associated, on the one hand, with a time marker and, on the other hand, with a cardiac electrogram track; a computer arranged to receive as input an electrogram track identifier and a time marker, to analyze the spatiotemporal dispersion signal of cardiac electrograms associated with this electrogram track identifier, and to analyze a signal extract between the time marker and the first preceding time marker whose cardiac electrogram spatiotemporal dispersion signal value indicates no dispersion, extract from this signal extract a signal flatness value and a duration value derived from the duration of the signal extract, and return a track priority value calculated from the flatness value and the duration value, and a monitor arranged to receive cardiac electrogram spatiotemporal dispersion signals associated with each cardiac electrogram track, and, when the value of a cardiac electrogram spatiotemporal dispersion signal indicates relevant dispersion, to call the calculator with the corresponding time marker and electrogram track identifier.

[0011] This device is particularly advantageous because it provides additional information beyond the dispersion value, indicating to the practitioner the priority level of the cardiac area responsible for the returned dispersion value in relation to fibrillation. Thus, the track priority value allows the practitioner to classify cardiac areas for ablation purposes.

[0012] Furthermore, the trail priority value tends to increase when a practitioner stops in a given area and that area contains a dispersion. Thus, the trail priority value allows the practitioner to perform their procedure more effectively: when a dispersion is detected, they know it is advisable to stop in the affected area and wait to see if the trail priority value increases. If it does not, they can be confident that it is advisable to continue their examination.

[0013] According to various embodiments, the invention may have one or more of the following characteristics: - Cardiac electrogram spatiotemporal dispersion signals are sequences of values ​​taken from electrogram signals, each indicating a degree of confidence that dispersion is occurring for the relevant electrogram track and time marker, and in which the monitor and / or calculator are arranged to determine whether a cardiac electrogram spatiotemporal dispersion signal indicates relevant dispersion by comparing the value of the cardiac electrogram spatiotemporal dispersion signal to a threshold value, - the calculator is configured to calculate the flatness value from at least one value among the standard deviation of the signal extract, the total variation of the signal extract, the entropy of the signal extract, or a value derived from one or more derivatives of order greater than or equal to one of the signal extract, - The calculator is designed to calculate the duration value by comparing the duration of the signal extract to a minimum and / or a maximum value, and returning the value 0 if the duration of the signal extract is less than the minimum value, by re- turning the value 1 if the duration of the signal extract is greater than the maximum value, and returning a value between 0 and 1 otherwise, - the calculator is configured to determine the value between 0 and 1 by applying to the duration of the signal extract a projection function from the interval between the minimum and maximum values ​​to the interval between 0 and 1, which projection function being chosen from the group comprising affine functions, exponential functions, polynomials and threshold functions, and - the computer receives spatiotemporal dispersion signals of cardiac electrograms associated on the one hand with a time marker and on the other hand with a cardiac electrogram track, b) determine if a value of a spatiotemporal dispersion signal of cardiac electrograms indicates dispersion, c) if operation b) is negative, repeat it with a dispersion value presenting a subsequent time marker, d) if operation b) is positive, analyze the spatiotemporal dispersion signal of cardiac electrograms associated with the corresponding electrogram track identifier by analyzing a signal extract between the time marker and the first preceding time marker whose spatiotemporal dispersion signal value of cardiac electrograms indicates an absence of dispersion, and by deriving from this signal extract a signal flatness value,and on the other hand a duration value derived from the duration of the signal extract, and e) return a track priority value calculated from the flatness value and the duration value of operation d).

[0014] Depending on various variations, the process may exhibit one or more of the following characteristics:

[0015] - spatiotemporal dispersion signals of cardiac electrograms are sequences of values ​​taken from electrogram signals, each indicating a degree of confidence that dispersion is occurring for the electrogram track and time marker concerned, and wherein operation b) comprises comparing a value of the spatiotemporal dispersion signal of cardiac electrograms to a threshold value, - operation d) includes calculating the flatness value from at least one value among the standard deviation of the signal extract, the total variation of the signal extract, the entropy of the signal extract, or a value derived from one or more derivatives of order greater than or equal to one of the signal extract, - Operation d) includes calculating the duration value by comparing the duration of the signal extract to a minimum and / or a maximum value, and returning the value 0 if the duration of the signal extract is less than the minimum value, by re- turning the value 1 if the duration of the signal extract is greater than the maximum value, and returning a value between 0 and 1 otherwise, - operation d) includes determining the value between 0 and 1 by applying to the duration of the signal extract a projection function from the interval between the minimum and maximum values ​​to the interval between 0 and 1, which projection function being chosen from the group comprising affine functions, exponential functions, polynomials and threshold functions, and - operation e) includes calculating the track priority value by performing a weighted harmonic, harmonic, weighted or arithmetic mean of the flatness value and the duration value.

[0016] The invention also relates to a computer program comprising instructions for executing the process according to the invention, a data storage medium on which such a computer program is recorded, and a computer system comprising a processor coupled to a memory, the memory having recorded such a computer program.

[0017] Other features and advantages of the invention will become clearer upon reading the following description, drawn from illustrative and non-limiting examples taken from the drawings shown: - Figure 1 represents a schematic diagram of a device according to the invention, and - [Fig.2] represents an example of the implementation of an operating loop of the device of [Fig.1],

[0018] The drawings and description below contain, for the most part, elements of a definite nature. They can therefore not only serve to better explain the present invention, but also contribute to its definition, if necessary.

[0019] Figure 1 represents a schematic example of a device 2 according to the invention. As stated in the introduction, the signals used by the device are based on cardiac electrograms measured by pairs of electrodes on a catheter in a patient's heart.

[0020] However, in the specific case of the invention, it is not these signals that are processed, but the dispersion measurement derived from them. As mentioned in the introduction, the article by Seitz et al., "AF Ablation Guided by Spatiotemporal Electrogram Dispersion Without Pulmonary Vein Isolation: A Wholly Patient-Tailored Approach", Journal of the American College of Cardiology, Volume 69, Issue 3, 24 January 2017, pages 303-321, and the patent published under number EP 3 236 843, provide a better understanding of what dispersion is, both in terms of similarities and differences with CFAE.

[0021] Given the field under consideration and the fact that dispersion and CFAE represent quite distinct phenomena, insofar as CFAE ignores everything From a spatial perspective, these values ​​are not of major interest within the scope of the invention. Indeed, as will be seen below, the invention generally aims to characterize the stability of the signal formed by the dispersion values. This same analysis would have little or no relevance in the case of CFAEs.

[0022] Device 2 includes a memory 4, a monitor 6 and a calculator 8.

[0023] Memory 4 is configured to receive all data from device 2, whether input or output, global or local. Memory 4 can be any type of data storage suitable for receiving digital data: hard drive, flash memory hard drive, flash memory of any kind, RAM, magnetic disk, locally distributed or cloud-based storage, etc.

[0024] In the example described here, memory 4 receives all the data pertaining to device 2, namely the programs and software instantiating monitor 6 and computer 8, their parameters and any hyperparameters, the weights of any neural networks, the outputs and intermediate data of the neural networks, the spatiotemporal dispersion signal data of cardiac electrograms received as input (if applicable), the signal flatness and duration values, the buffered data, and the output track priority value data. The data computed by the device can be stored on any type of memory similar to memory 4, or on memory 4 itself. This data can be erased after the device has completed its tasks or retained.

[0025] As will be seen below, the signal flatness value and the duration value are two values ​​used to determine whether the signal from dispersion values ​​is associated with a cardiac area for which ablation is a priority. These values ​​are combined to produce a track priority value, which indicates whether the determined dispersion value is associated with a dispersion that has a significant potential to be associated with an area that is the source of atrial fibrillation. Even more so than in the Applicant's other patent applications, the track priority value is not a diagnosis but an indication that allows a physician to make a decision, much like blood pressure would be in another context.

[0026] Monitor 6 and calculator 8 access memory 4 directly or indirectly. They can be implemented as suitable computer code running on one or more processors. By processors, we mean any processor suitable for the calculations described below. Such a processor can be implemented in any known way, such as a microprocessor for a personal computer, laptop, tablet, or smartphone; a dedicated chip of the FPGA or SoC type; a computing resource on a grid or in the cloud; a graphics processing unit (GPU) array; a microcontroller; or any other form suitable for providing the The computing power required for the implementation described below. One or more of these elements can also be implemented as specialized electronic circuits such as an ASIC. A combination of processor and electronic circuits can also be considered. Processors dedicated to machine learning could also be considered.

[0027] Monitor 6 is designed to analyze the input spatiotemporal dispersion signal data stream of cardiac electrograms and detect when a dispersion value indicates that treatment is necessary. As mentioned in the introduction, within the VX1 software, a dispersion value greater than or equal to 0.5 is considered significant. Of course, the detection of this value will depend on the values ​​of the input dispersion signal. For example, it could be generated inversely by the VX1 software (e.g., 1 - Value from VX1), in which case a value less than or equal to 0.5 would be significant. This determination could also be made differently, based on a value derived from the derivative of the dispersion signal, or in any other relevant manner.

[0028] Due to the continuous nature of the processing by device 2, which will become clearer below, once a dispersion value has been detected by monitor 6, the detection for subsequent values ​​(but concerning the same track, of course) can be different or simplified. Thus, in the case described above, rather than comparing the current dispersion value to a threshold, monitor 6 could, for example, measure the derivative of the input dispersion value signal and consider a detection positive if the derivative is positive, and so on. In general, monitor 6 could rely on several tests to qualify the detection of a relevant dispersion value.

[0029] The role of monitor 6 is therefore "interruptive." Indeed, in the absence of a relevant dispersion value, it is pointless to calculate a track priority value, since no dispersion is detected. Conversely, as soon as a relevant dispersion value is detected, monitor 6 calls calculator 8 to calculate the track priority value. Thus, the operation of device 2 can be seen as a detection loop by monitor 6 for each track, with calculator 8 executing each time a relevant dispersion value is detected. Of course, other implementations could be considered.

[0030] Calculator 8 calculates the track priority value for a track whose dispersion value has been deemed relevant by monitor 6. As a reminder, the dispersion signal is a signal that associates a time marker with a dispersion value. This dispersion value is itself derived from an analysis of several cardiac electrogram signal values. Within the software VX1, dispersion values ​​are updated approximately every 300ms, based on extracts of cardiac electrogram signals lasting approximately 1.5s. Alternatively, this update can occur every 100ms, every 500ms, or other intervals.

[0031] As we will see below, the determination of a track priority value is based on an extract of dispersion values ​​which can have a continuous duration of 1.5s, i.e. about 5 dispersion values, up to several tens of seconds, i.e. about a hundred dispersion values.

[0032] More specifically, the calculator 8 analyzes a signal extract of dispersion values ​​each time, ending with the dispersion value that monitor 6 has just determined to be relevant. This extract contains exclusively dispersion values ​​whose time markers are consecutive, associated with the same track, and considered relevant by monitor 6. It will become apparent that this extract can be obtained in numerous ways: - Monitor 6 can generate extracts during its operation, by adding a current dispersion value detected as relevant to a current extract if the immediately preceding dispersion value was also detected as relevant, or creating a new extract otherwise. - Calculator 8, upon receiving a dispersion value associated with a given time marker, can analyze a buffer of past dispersion values ​​and stop at the oldest value considered relevant by monitor 6, or alternatively, - Calculator 8 can retrieve a buffer of past dispersion values ​​from the time marker associated with an input received dispersion value, and properly determine an extract from this buffer that it considers relevant.

[0033] The Applicant's work has demonstrated that the temporal continuity of relevant dispersion values ​​is all the more useful when the dispersion value signal is precise. Indeed, with a dispersion considered "noisy," one might be tempted to ignore an irrelevant dispersion value in order to obtain more data for calculating the track priority value. The Applicant's work has demonstrated that combining the signal flatness value and the duration value avoids artificially increasing the size of the extracts and yields better results.

[0034] Calculator 8 operates by performing two measurements on the extract defined above: a measurement of the signal's flatness value and a measurement of its duration value. In both cases, the aim is to determine whether the dispersion value exhibits some form of stability over time. Indeed, the Applicant's work has revealed that signals with stable dispersion values ​​are associated with priority areas for ablation in order to eliminate atrial fibrillation.

[0035] In the example described here, the signal flatness value is derived from the standard deviation of the data from the extract. The Applicant's work revealed that the standard deviation is the measure providing the best results. Nevertheless, the Applicant determined that other types of measures could be retained, such as variance, total variation of the extract, entropy of the extract, a value taken from one or more derivatives of order greater than or equal to one of the extract, range, interquartile range or another similar measure.

[0036] In parallel, Calculator 8 also determines a duration value that indicates how long the current extract is relative to a time interval considered to indicate that the cardiac area associated with this dispersion has priority for ablation. In the example described here, Calculator 8 projects the extract duration onto a standard interval between a minimum and a maximum duration. These two values, empirically estimated by the Applicant, respectively indicate the minimum duration an extract must have to designate a priority cardiac area, and the maximum duration, given that the practitioner cannot afford to remain too long on each area if they wish to perform the procedure within reasonable timeframes and minimize the operative risk.

[0037] In the example described here, calculator 8 determines the duration of the extract and projects its value onto the interval [minimum duration; maximum duration] to determine a value between 0 and 1. The projection can be of any type: linear, polynomial, exponential, threshold, etc. The aim here is to indicate, for a given extract duration, whether this duration is characteristic of a priority cardiac zone or not.

[0038] The minimum duration has a clear utility as a floor value. The maximum duration has an important operational utility: when the first dispersion value of an extract longer than the maximum duration is detected, the output priority value will necessarily be low because the duration value will be small. As the extract grows, the track priority value will increase with the duration value. When an operator observes that the track priority value is no longer changing, they can determine that this is because the duration value can no longer increase, and that it is time to move the catheter.

[0039] Alternatively, calculator 8 could determine the duration value differently, independently of the range [minimum duration; maximum duration]. Also alternatively, the minimum and maximum durations could vary during the procedure or be customized for each patient.

[0040] The signal flatness value and the duration value can be determined in parallel. Alternatively, one can be calculated before the other.

[0041] The value between 0 and 1 is chosen because of how the calculator 8 determines the track priority value from the signal flatness value and the duration value. Indeed, the calculator 8 operates in the example described by operating a weighted harmonic mean. Alternatively, this mean can be harmonic, weighted or arithmetic of the flatness value and the duration value.

[0042] In the event that the signal flatness value and the duration value are not within identical ranges, a relative rectification may be performed, or another formula may be used to calculate the track priority value.

[0043] Figure 2 illustrates an example of the operating loop of device 2. In operation 200, monitor 6 is called with the current dispersion value for a given track. When monitor 6 determines that a dispersion value is relevant, operation 210 is initiated in which the extract is determined from the time marker of the dispersion value in operation 200, the associated track identifier, and any dispersion value signals already received for that track identifier. Once the extract is determined, the computer 8 can determine the flatness value and the duration value in operation 220, and then the track priority value and return this value in operation 230.

[0044] It appears that, where applicable, device 2 will determine a track priority value for each track for which a dispersion value signal is received as input. This enriches the information transmitted to the practitioner, enabling them to make the diagnosis that determines whether an area associated with a given track should be ablated or not.

Claims

Claims

1. Device for processing spatiotemporal dispersion signals of cardiac electrograms comprising a memory (4) arranged to receive spatiotemporal dispersion signals of cardiac electrograms associated on the one hand with a time marker and on the other hand with a cardiac electrogram track, a computer (8) arranged to receive as input an electrogram track identifier and a time marker, to analyze the spatiotemporal dispersion signal of cardiac electrograms associated with this electrogram track identifier, analyze a signal extract between the time marker and the first preceding time marker whose spatiotemporal dispersion signal value of cardiac electrograms indicates an absence of dispersion, derive from this signal extract on the one hand a signal flatness value, and on the other hand a duration value derived from the duration of the signal extract,and returning a track priority value calculated from the flatness value and the duration value, and a monitor (6) arranged to receive the cardiac electrogram spatiotemporal dispersion signals associated with each cardiac electrogram track, and, when the value of a cardiac electrogram spatiotemporal dispersion signal indicates a relevant dispersion, to call the calculator (8) with the corresponding time marker and electrogram track identifier.,

2. Apparatus according to claim 1, wherein the cardiac electrogram spatiotemporal dispersion signals are sequences of values ​​derived from electrogram signals each indicating a degree of confidence that dispersion is occurring for the relevant electrogram track and time marker, and wherein the monitor (6) and / or the calculator (8) are arranged to determine whether a cardiac electrogram spatiotemporal dispersion signal indicates relevant dispersion by comparing the value of the cardiac electrogram spatiotemporal dispersion signal with a threshold value.

3. Device according to claim 1 or 2, in which the calculator (8) is arranged to calculate the flatness value from at least one value among the standard deviation of the signal extract, the total variation of the signal extract, the entropy of the signal extract or a value taken from one or more derivatives of order greater than or equal to one of the signal extract. signal.

4. Device according to one of the preceding claims, in which the calculator (8) is arranged to calculate the duration value by comparing the duration of the signal extract to a minimum value and / or to a maximum value, and returning the value 0 if the duration of the signal extract is less than the minimum value, returning the value 1 if the duration of the signal extract is greater than the maximum value, and returning a value between 0 and 1 otherwise.

5. Device according to claim 4, in which the calculator (8) is arranged to determine the value between 0 and 1 by applying to the duration of the signal extract a projection function of the interval between the minimum value and the maximum value towards the interval between 0 and 1, which projection function is chosen from the group comprising affine functions, exponential functions, polynomials and threshold functions.

6. Device according to one of the preceding claims, in which the calculator (8) is arranged to calculate the track priority value by performing a weighted harmonic, harmonic, weighted or arithmetic mean of the flatness value and the duration value.

7. A method for determining a track priority value of cardiac electrogram spatiotemporal dispersion signals comprising the following steps: a) receiving cardiac electrogram spatiotemporal dispersion signals associated on the one hand with a time marker and on the other hand with a cardiac electrogram track, b) determining whether a value of a cardiac electrogram spatiotemporal dispersion signal indicates dispersion, c) if step b) is negative, repeating it with a dispersion value having a later time marker, d) if step b) is positive,analyzing the spatio-temporal dispersion signal of cardiac electrograms associated with the corresponding electrogram track identifier by analyzing a signal extract between the time marker and the first preceding time marker whose spatio-temporal dispersion signal value of cardiac electrograms indicates an absence of dispersion, and by deriving from this signal extract on the one hand a signal flatness value, and on the other hand a duration value derived from the duration of the signal extract, and, e) return a track priority value calculated from the flatness value and the duration value of operation d).

8. A method according to claim 7, wherein the cardiac electrogram spatiotemporal dispersion signals are sequences of values ​​derived from electrogram signals each indicating a degree of confidence that dispersion is occurring for the electrogram track and time marker concerned, and wherein step b) comprises comparing a value of the cardiac electrogram spatiotemporal dispersion signal to a threshold value.

9. The method of claim 7 or 8, wherein step d) comprises calculating the flatness value from at least one of the standard deviation of the signal extract, the total variation of the signal extract, the entropy of the signal extract, or a value derived from one or more derivatives of order greater than or equal to one of the signal extract.

10. Method according to one of claims 7 to 9, in which operation d) comprises calculating the duration value by comparing the duration of the signal extract to a minimum value and / or a maximum value, and returning the value 0 if the duration of the signal extract is less than the minimum value, returning the value 1 if the duration of the signal extract is greater than the maximum value, and returning a value between 0 and 1 otherwise.

11. Method according to claim 10, in which operation d) comprises determining the value between 0 and 1 by applying to the duration of the signal extract a projection function of the interval between the minimum value and the maximum value towards the interval between 0 and 1, which projection function is chosen from the group comprising affine functions, exponential functions, polynomials and threshold functions.

12. A method according to one of claims 7 to 11, wherein operation e) comprises calculating the track priority value by performing a weighted harmonic, harmonic, weighted or arithmetic mean of the flatness value and the duration value.

13. A computer program comprising instructions for carrying out the method according to one of claims 7 to 12 when executed by computer.

14. Data storage medium on which the computer program according to claim 13 is recorded.

Images