Cardiac mapping system with efficiency algorithm

Abstract

A dynamic display system of cardiac information includes one or more electrodes configured to record a plurality of time intervals of a set of electrical potential data representative of cardiac activity, and a cardiac information console including a signal processor configured to calculate a plurality of time intervals of a set of cardiac activity data using the recorded set of electrical potential data, wherein the cardiac activity data is associated with surface locations of one or more chambers of the heart, and a user interface module configured to display a series of images, each image including a graphical representation of cardiac activity. Methods of providing cardiac activity data are also provided.

Description

[0001] Relevant application data

[0002] This application claims priority to U.S. Provisional Patent Application No. 62 / 413,104, entitled “Cardiac Mapping System with Efficiency Algorithm,” filed October 26, 2016, and U.S. Provisional Patent Application No. 62 / 331,364, entitled “Cardiac Mapping System with Efficiency Algorithm,” filed May 3, 2016, each of which is incorporated herein by reference in its entirety.

[0003] While this application does not claim priority to the following, it may be related to: U.S. Patent Application No. 14 / 865,435, filed September 25, 2015, entitled "Method and Apparatus for Determining and Presenting Surface Charge and Dipole Density on a Heart Wall," which is a continuation of U.S. Patent No. 9,167,982, filed November 19, 2014, entitled "Method and Apparatus for Determining and Presenting Surface Charge and Dipole Density on a Heart Wall," which is a continuation of U.S. Patent No. 8,918,158 (hereinafter referred to as the '158 Patent), published December 23, 2014, entitled "Method and Apparatus for Determining and Presenting Surface Charge and Dipole Density on a Heart Wall." The continuation of U.S. Patent No. 8,700,119 (hereinafter referred to as the '119 Patent), entitled "Method and Apparatus for Determining and Presenting Surface Charge and Dipole Density on the Heart Wall," published on April 15, 2014, is a continuation of U.S. Patent No. 8,417,313 (hereinafter referred to as the '313 Patent), entitled "Method and Apparatus for Determining and Presenting Surface Charge and Dipole Density on the Heart Wall," published on April 9, 2013, which is a continuation of PCT application CH2007 / 000380, entitled "Method and Apparatus for Determining and Presenting Surface Charge and Dipole Density on the Heart Wall," filed on August 3, 2007, and claims priority to Swiss Patent Application No. 1251 / 06, filed on August 3, 2006, each of which is incorporated herein by reference.

[0004] While this application does not claim priority to the following, it may be related to: U.S. Patent Application No. 14 / 886,449, filed October 19, 2015, entitled “Apparatus and Method for Geometrically Determining Electric Dipole Density on a Heart Wall,” which is a continuation of U.S. Patent No. 9,192,318, filed July 19, 2013, entitled “Apparatus and Method for Geometrically Determining Electric Dipole Density on a Heart Wall,” published August 20, 2013. This is a continuation of U.S. Patent No. 8,512,255, entitled “Apparatus and Method for Geometrically Determining Electric Dipole Density on a Core Wall,” published as US2010 / 0298690 (hereinafter referred to as '690 Publication). It is a 35 USC 371 national phase application filed on January 16, 2009, entitled “Apparatus and Method for Geometrically Determining Electric Dipole Density on a Core Wall,” under Patent Cooperation Treaty Application PCT / IB09 / 00071, which claims priority to Swiss Patent Application No. 00068 / 08, filed on January 17, 2008, each of which is incorporated herein by reference.

[0005] While this application does not claim priority to the following, it may be related to: U.S. Application 14 / 003671, filed September 6, 2013, entitled “Apparatus and Method for Geometrically Determining Electric Dipole Density on a Core Wall,” which is a Patent Cooperation Treaty application PCT / US2012 / 028593, entitled “Apparatus and Method for Geometrically Determining Electric Dipole Density on a Core Wall,” published as WO2012 / 122517 (hereinafter referred to as '517 Publication), which claims priority to U.S. Provisional Patent Application 61 / 451,357, each of which is incorporated herein by reference.

[0006] While this application does not claim priority to the following, it may be related to: U.S. Design Application No. 29 / 475,273, filed December 2, 2013, entitled “Catheter System and Method for Medical Use Thereof, Including Diagnostic and Therapeutic Use in the Heart”; National Phase 35 USC 371 of Patent Cooperation Treaty Application PCT / US2013 / 057579, filed August 30, 2013, entitled “Catheter System and Method for Medical Use Thereof, Including Diagnostic and Therapeutic Use in the Heart”, which claims priority to U.S. Provisional Patent Application No. 61 / 695,535, filed August 31, 2012, entitled “System and Method for Diagnosing and Treating Cardiac Tissue”, each of which is incorporated herein by reference.

[0007] While this application does not claim priority to the following, it may be related to Patent Cooperation Treaty application PCT / US2014 / 15261, filed February 7, 2014, entitled "Expandable Conduit Assembly with Flexible Printed Circuit Board (PCB) Electrical Paths," which claims priority to U.S. Provisional Patent Application No. 61 / 762,363, filed February 8, 2013, entitled "Expandable Conduit Assembly with Flexible Printed Circuit Board (PCB) Electrical Paths," each of which is incorporated herein by reference.

[0008] While this application does not claim priority to the following, it may be related to: Patent Cooperation Treaty application PCT / US2015 / 11312, entitled “Gas-out Patient Access Device,” filed January 14, 2015, which claims priority to U.S. Provisional Patent Application No. 61 / 928,704, entitled “Gas-out Patient Access Device,” filed January 17, 2014, each of which is incorporated herein by reference.

[0009] While this application does not claim priority to the following, it may be related to: Patent Cooperation Treaty application PCT / US2015 / 22187, filed March 24, 2015, entitled “Heart Analysis User Interface System and Method,” which claims priority to U.S. Provisional Patent Application No. 61 / 970,027, filed March 28, 2014, entitled “Heart Analysis User Interface System and Method,” each of which is incorporated herein by reference.

[0010] While this application does not claim priority to the following, it may be related to Patent Cooperation Treaty application PCT / US2014 / 54942, filed September 10, 2014, entitled “Apparatus and method for determining electric dipole density on a cardiac surface,” which claims priority to U.S. Provisional Patent Application 61 / 877,617, filed September 13, 2013, entitled “Apparatus and method for determining electric dipole density on a cardiac surface,” each of which is incorporated herein by reference.

[0011] While this application does not claim priority to the following, it may be related to: U.S. Provisional Patent Application No. 62 / 161,213, filed May 13, 2015, entitled “Location System and Method for Acquiring and Analyzing Cardiac Information,” which is incorporated herein by reference.

[0012] While this application does not claim priority to the following, it may be related to: U.S. Provisional Patent Application No. 62 / 160,501, filed May 12, 2015, entitled “Cardiac Virtualization Test Box and Testing System and Method,” which is incorporated herein by reference.

[0013] While this application does not claim the following priority, it may be related to: U.S. Provisional Patent Application No. 62 / 160,529, entitled “Ultrasound Sequencing System and Method,” filed May 12, 2015, which is incorporated herein by reference. Technical Field

[0014] This invention generally relates to systems and methods that may be used for diagnosing and treating arrhythmias or other abnormalities, and in particular, to systems, apparatus, and methods for displaying cardiac activity associated with the diagnosis and treatment of these arrhythmias or other abnormalities. Background Technology

[0015] Cardiac signals (e.g., charge density, dipole density, voltage, etc.) vary in amplitude across the endocardial surface. The amplitude of these signals depends on several factors, including local tissue characteristics (e.g., health vs. disease / scarring / fibrosis / injury) and regional activation characteristics (e.g., the “electrical mass” of activated tissue prior to the activation of local cells). A common practice is to assign a single threshold to all signals across the surface at all times. Using a single threshold can lead to missing low-amplitude activations or allowing high-amplitude activations to dominate / saturate, resulting in misinterpretation of the images. Failure to properly detect activation can lead to inaccurate identification of regions of interest for treatment delivery or incomplete assessment of ablation efficacy (over- or under-blocking).

[0016] Continuous global mapping of atrial fibrillation (AF) yields a large number of temporally and spatially variable activation patterns. Limited discrete sampling of image data may be insufficient to provide a complete picture of the drivers, mechanisms, and supporting substrates of the arrhythmia. Long-term clinical review of AF may be difficult to remember and piece together into a “larger picture.” Summary of the Invention

[0017] According to one aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: an automatic orientation system configured to orient one or more positioning axes in a desired direction and / or rotate the positioning coordinate system toward a physiological orientation.

[0018] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: a cardiac information algorithm that directly performs truncated SVD calculations to calculate the set of cardiac activity data.

[0019] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: a non-functional electrode detection algorithm configured to perform sufficient functional evaluation of one or more electrodes.

[0020] In some embodiments, the signal processor is also configured to exclude voltages measured from electrodes identified by a non-functional electrode detection algorithm from the calculation of the cardiac activity data set.

[0021] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the console includes: an analog-to-digital converter; and a subsystem configured to automatically shut down the analog-to-digital converter under certain conditions.

[0022] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: an algorithm configured to reduce the influence of far-field signals of the ventricles.

[0023] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: a registration processing subsystem configured to associate voltage measurements with locations.

[0024] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: an improved field estimation algorithm.

[0025] According to another aspect of the present invention, a dynamic display system for cardiac information includes: one or more electrodes configured to record a set of potential data representing cardiac activity at multiple time intervals; and a cardiac information console including: a signal processor configured to calculate a set of cardiac activity data at multiple time intervals using the recorded set of potential data, wherein the cardiac activity data is associated with surface locations of one or more cardiac chambers; and a user interface module configured to display a series of images, each image including: a graphical representation of cardiac activity; wherein the signal processor includes: a forward matrix and an inverse solution; and a scaling algorithm including a pre-scaling portion configured to reduce the distance effect on the forward matrix, and a post-scaling portion for adjusting the output of the inverse solution. Attached Figure Description

[0026] Figure 1This is a block diagram of an embodiment of a cardiac information processing system according to an aspect of the present invention.

[0027] Figure 1-1 (a) and Figure 1-1 (b) are two flowcharts illustrating a method for performing SVD calculations according to an aspect of the present invention.

[0028] Figure 2 These are front and rear views of the patient and the relative electrode placement according to an aspect of the present invention.

[0029] Figure 3 This is a flowchart illustrating the identification of non-functional electrodes according to an aspect of the present invention.

[0030] Figure 4 This is a flowchart illustrating the V-wave removal process according to an aspect of the present invention.

[0031] Figures 4-1 to 4-3 Various examples of V-wave removal curves based on aspects of the present invention are shown.

[0032] Figure 5 This is a flowchart of an algorithm for performing localization according to an aspect of the present invention.

[0033] Figure 6 It is a 3D display of the integrated path to the target frame according to aspects of the present invention.

[0034] Figure 7 It is a 3D display of the integrated path to the target system according to aspects of the present invention.

[0035] Figure 8 This is a flowchart of a scaling algorithm based on an aspect of the present invention.

[0036] Figure 9 A flowchart illustrating a method for mapping the positions of multiple baskets according to an aspect of the present invention is shown.

[0037] Figure 10 A flowchart illustrating the steps for reclassifying non-functional electrodes for calculating cardiac activity, according to an aspect of the present invention, is shown.

[0038] Figure 11A and 11B Two examples of V-wave removal (VWR) are shown.

[0039] Figure 12 This is an embodiment of a display of a view of cardiac activation data according to aspects of the present invention.

[0040] Figure 13Examples of applying a median filter, based on aspects of the present invention, to propagate historical images are shown.

[0041] Figure 14 An exemplary implementation of false positive activation annotation removal according to an aspect of the present invention is shown.

[0042] Figure 15 A lasso-type catheter, based on an aspect of the invention, is shown.

[0043] Figure 16 This is a flowchart illustrating a method for fitting points to a curve representing a duct, according to an aspect of the invention.

[0044] Figures 17A to 17C An example of a pace blanking method for one or more channels, based on aspects of the present invention, is shown.

[0045] Figure 18A and 18B Two examples of the use of voxels according to aspects of the present invention are shown.

[0046] Figure 19 This is a flowchart of an embodiment of a method for generating a beating anatomical model of the heart, according to an aspect of the present invention.

[0047] Figure 20 This is a flowchart of an embodiment of a method for displaying a dynamic model of the heart, according to an aspect of the present invention.

[0048] Figure 21 This is a flowchart of an embodiment of a method for editing a heart model according to aspects of the present invention. Detailed Implementation

[0049] Referring to the accompanying drawings, various illustrative embodiments will be described more fully below, some of which are illustrated in the drawings. However, the inventive concept can be embodied in different forms and should not be construed as limiting the exemplary embodiments set forth herein.

[0050] It should be understood that although the terms first, second, etc., are used herein to describe various elements, these elements are not limited by these terms. The use of these terms to distinguish one element from another does not imply a desired sequence of elements. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the invention. As used herein, the term "and / or" includes any and all combinations of one or more related listed items. Furthermore, a "combination" of related listed items does not need to include all listed items, but may include all listed items.

[0051] It should be understood that when one element is “on” or “attached,” “connected,” or “coupled” to another element, it can be directly on or connected to the other element, or there may be an intermediate element. Conversely, when an element is described as being “directly on” or “directly connected” or “directly coupled” to another element, there is no intermediate element. Other terms used to describe relationships between elements should be interpreted in a similar manner (e.g., “between” vs. “directly between”, “adjacent” vs. “directly adjacent,” etc.).

[0052] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein are intended to include the plural forms as well. It will be further understood that when the terms “comprising,” “including,” and / or “containing” are used herein, the terms specifically refer to the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0053] Spatial relation terms, such as “below,” “under,” “below,” “above,” “over,” etc., may be used to describe the relationship of one element and / or feature to another element and / or feature, for example, as shown in the figures. It will be understood that, in addition to the orientations depicted in the figures, spatial relation terms are intended to encompass different orientations of the device during use and / or operation. For example, if the device in the figures is flipped, an element or feature described as “below” and / or “under” will be oriented “above” other elements or features. The device may be oriented in other ways (e.g., rotated 90 degrees or in other orientations), and the spatial relative descriptors used herein are interpreted accordingly.

[0054] Localization describes the process of establishing a coordinate system and using one or more signals (e.g., electronic signals) to determine the location of one or more objects within that system. In some embodiments, the localization process includes one or more signals generated from one or more sources that change spatially and / or temporally, and a sensor, detector, or other transducer that measures the generated signals from a location. The sensor may be located on or separate from the object being localized. Analysis and / or calculation of the measured signals can be used to determine the positional relationship between the sensor and / or object and one or more sources of the generated signals. Localization methods may include two or more generated signals to increase the number or accuracy of positional relationships between the sensor and the sources. The sources, sensors, and / or objects may be co-located or may be the same device. In some embodiments, changes in time and / or space include the interaction of the generated signals with the measurement environment. In other embodiments, the localization process measures inherent or existing properties or characteristics of the object, sensor, or environment, such as measuring signals from an accelerometer located on the object or sensor.

[0055] This document describes various exemplary embodiments with reference to idealized or representative structures and intermediate structures. Thus, variations in the example shapes are expected as a result of, for example, manufacturing techniques and / or tolerances. Therefore, the exemplary embodiments should not be construed as limiting the specific shapes of the areas shown herein, but rather include, for example, deviations in shape due to manufacturing processes.

[0056] Wherever the functional features, operations, and / or steps are described herein or otherwise understood to be included in the various embodiments of the inventive concept, these functional features, operations, and / or steps may be embodied in functional blocks, units, modules, operations, and / or methods. Furthermore, wherever these functional blocks, units, modules, operations, and / or methods include computer program code, such computer program code may be stored in a computer-readable medium, such as a non-transitory memory and medium executable by at least one computer processor.

[0057] Now for reference Figure 1 A block diagram of an embodiment of a cardiac information processing system 100 according to aspects of the present invention is provided. The cardiac information processing system 100 may be or include a system configured to perform cardiac mapping, diagnosis, and / or treatment, for example, for treating a patient with a disease or condition such as arrhythmia. Additionally or alternatively, the system may be a system configured to teach and / or validate means and methods for diagnosing and / or treating cardiac abnormalities or diseases in patient P. The system may also be used to generate a display of cardiac activity, such as a dynamic display of active wavefronts propagating on the surface of the heart.

[0058] The cardiac information processing system 100 includes a catheter 10, a cardiac information console 20, and a patient interface module 50, which can be configured to cooperate to perform various functions of the cardiac information processing system 100. The cardiac information processing system 100 may include a single power supply (PWR) that can be shared by the cardiac information console 20 and the patient interface module 50. Using a single power supply in this way can significantly reduce the chance of leakage current propagating into the patient interface module 50 and causing errors in positioning (i.e., the process of determining the location of one or more electrodes within the patient P). The cardiac information console 20 includes a bus 21 that electrically or operatively connects various components of the console 20 to each other, such as... Figure 1 As shown.

[0059] The catheter 10 includes an electrode array 12 that can be percutaneously delivered to a cardiac chamber (HC). In this embodiment, the electrode array 12 has a known spatial configuration in three-dimensional (3D) space. For example, the physical relationships of the electrode array 12 in an inflated state may be known or reliably assumed. The diagnostic catheter 10 also includes a handle 14 and an elongated flexible shaft 16 extending from the handle 14. Attached to the distal end of the shaft 16 is the electrode array 12, such as a radially expandable and / or compressible component. In this embodiment, the electrode array 12 is shown as a basket array, but the electrode array 12 may take other forms in other embodiments. In some embodiments, the expandable electrode array 12 may be constructed and arranged as described in the following applications of the applicant: International PCT patent application PCT / US2013 / 057579, filed August 30, 2013, entitled "System and method for diagnosing and treating cardiac tissue," and International PCT patent application PCT / US2014 / 015261, filed February 7, 2014, entitled "Expandable catheter assembly with flexible printed circuit board," the contents of which are incorporated herein by reference in their entirety. In other embodiments, the expandable electrode array 12 may include a balloon, radially deployable arms, a helical array, and / or other expandable and compressible structures (e.g., elastically biased structures).

[0060] The shaft 16 and the expandable electrode array 12 are configured and arranged to be inserted into a body (e.g., an animal or human body, such as the body of patient P) and advanced via a blood vessel (e.g., the femoral vein and / or other blood vessels). The shaft 16 and the electrode array 12 may be configured and arranged to be inserted via a guide (not shown), for example when the electrode array 12 is in a compressed state, and slide through the lumen of the guide into a body space, such as a chamber (HC) of the heart, such as the right or left atrium.

[0061] The expandable electrode array 12 may include multiple splines (e.g., elastically biased at...) Figure 1Multiple splines in the basket shape shown, each spline having multiple biopotential electrodes 12a and / or multiple ultrasonic (US) transducers 12b. Figure 1 Three splines can be seen in the array, but the basket array is not limited to three splines; it can include more or fewer splines. Each electrode 12a can be configured to record a bioelectric potential (or voltage), such as a voltage determined (e.g., measured or sensed) at a point on the surface of the heart or at a location within the heart chamber (HC). Each US transducer 12b can be configured to transmit ultrasound signals and receive ultrasound reflections to determine the range to a reflecting target, such as a point on the surface of the heart chamber (HC), which is used in the creation of digital models of anatomical structures.

[0062] As a non-limiting example, in this embodiment, three electrodes 12a and three US transducers 12b are shown on each spline. However, in other embodiments, the basket array may include more or fewer electrodes and / or more or fewer US transducers. Furthermore, electrodes 12a and transducers 12b may be arranged in pairs. Here, one electrode 12a is paired with one transducer 12b, and each spline has multiple electrode-transducer pairs. However, the inventive concept is not limited to this particular electrode-transducer arrangement. In other embodiments, not all electrodes 12a and transducers 12b need to be arranged in pairs; some may be arranged in pairs while others are not. Moreover, in some embodiments, not all splines need to have the same arrangement as the electrodes 12a and transducers 12b. Additionally, in some embodiments, electrodes 12a may be arranged on some splines, while transducers 12b may be arranged on other splines.

[0063] The catheter 10 may include cables or other conduits, such as cable 18, configured to electrically, optically, and / or electro-optically connect the catheter 10 to the cardiac information console 20 via connectors 18a and 20a, respectively. In some embodiments, cable 18 includes mechanisms selected from the group consisting of: cables, such as steering cables; mechanical connections; hydraulic lines; pneumatic lines; and combinations of one or more of these.

[0064] The patient interface module 50 can be configured to electrically isolate one or more components of the cardiac information console 20 from the patient P (e.g., to prevent unintended transfer of shock or other unwanted electrical energy to the patient P). The patient interface module 50 can be integrated with the cardiac information console 20, and / or it can comprise separate discrete components (e.g., separate housings), as shown. The cardiac information console 20 includes one or more connectors 20b, each including a jack, plug, terminal, port, or other custom or standard electrical connector, optical connector, and / or mechanical connector. In some embodiments, the connectors 20b are terminated to maintain the desired input impedance at different RF frequencies (e.g., 10 kHz to 20 MHz). In some embodiments, termination can be achieved by using a filter to terminate the cable shield. In some embodiments, the termination filter can provide high input impedance over a frequency range, for example, to minimize leakage at the positioning frequency, and low input impedance over different frequency ranges, for example, to achieve maximum signal integrity at ultrasound frequencies. Similarly, the patient interface module 50 includes one or more connectors 50b. At least one cable 52 connects the patient interface module 50 to the cardiac information console 20 via connectors 20b and 50b.

[0065] In this embodiment, the patient interface module 50 includes an isolated positioning drive system 54, a set of patch electrodes 56, and one or more reference electrodes 58. The isolated positioning drive system 54 isolates the positioning signal from the rest of the system to prevent current leakage, such as signal loss, which could lead to performance degradation. In some embodiments, the isolation of the positioning signal from the rest of the system includes an impedance range greater than 100 kiloohms, for example, approximately 500 kiloohms at the positioning frequency. The isolation of the positioning drive system 54 minimizes positional drift and maintains a high degree of isolation between axes. The positioning drive system 54 can operate as a current, voltage, magnetic, acoustic, or other type of energy mode driver. Based on the energy mode employed by the positioning drive system 54, the set of patch electrodes 56 and / or one or more reference electrodes 58 can consist of conductive electrodes, magnetic coils, acoustic transducers, and / or other types of transducers or sensors. Furthermore, the isolated positioning drive system 54 maintains simultaneous output on all axes, such as positioning signals present on each axis electrode pair, while also increasing the effective sampling rate at each electrode location. In some embodiments, the location sampling rate includes a rate from 10 kHz to 20 MHz, such as a sampling rate of approximately 625 kHz.

[0066] In this embodiment, the patch electrode group 56 includes three (3) pairs of patch electrodes: an "X" pair having two patch electrodes (X1, X2) placed on opposite sides of the ribs; a "Y" pair having one patch electrode (Y1) placed on the lower back and one patch electrode (Y2) placed on the upper chest; and a "Z" pair having one patch electrode (Z1) placed on the upper back and one patch electrode (Z2) placed on the lower abdomen. The patch electrode pairs 56 can be placed on any orthogonal and / orthogonal axis group. Figure 1 In one embodiment, the placement of electrodes on patient P is shown, with electrodes on the back shown in dashed lines. (See also...) Figure 2 )

[0067] Figure 2 These are front and rear views of the patient P and the relative electrode placement according to an aspect of the invention. The figure illustrates a preferred patch electrode placement, as described above. Figure 1 In this diagram, for example, X electrodes X1 and X2 are shown as patch electrodes 1 and 2, respectively; Z electrodes Z1 and Z2 are shown as patch electrodes 3 and 4, respectively; and Y electrodes Y1 and Y2 are shown as patch electrodes 5 and 6, respectively. Therefore, patches 1 and 2 are placed on the ribs, forming the X-axis within the body; patches 3 and 4 are placed on the lower back and upper chest, respectively, forming the Z-axis; and patches 5 and 6 are placed on the upper back and lower abdomen (torso), respectively, forming the Y-axis. The three axes have similar lengths and are not aligned with the body's "natural" axes (i.e., from head to toe, from chest to back, and from side to side).

[0068] Reference patch electrode 58 can be placed on the lower back / buttocks. Additionally or alternatively, a reference catheter can be placed within a blood vessel in the body, such as in or near a blood vessel in the lower back / buttocks.

[0069] The placement of electrodes 56 defines a coordinate system consisting of three axes, with one axis for each pair of patch electrodes 56. In some embodiments, such as... Figure 2 As shown, the axes are not orthogonal to the body's natural axes, such as those from head to toe, from chest to back, and from side to side (i.e., from rib to rib). The electrodes can be positioned such that the axes intersect at an origin, such as the origin within the heart. For example, the origin of the three intersecting axes could be centered within the volume of the atria. System 100 can be configured to provide an "electrical zero" located outside the heart, for example, by positioning reference electrode 58 such that the resulting electrical zero is outside the heart (e.g., to avoid a transition from positive to negative voltage at one or more of the positioned locations).

[0070] As described above, the patch pairs can operate differentially, meaning that none of the patches 56 in the pair operate as reference electrodes, and both are driven by system 100 to generate an electric field between them. Alternatively or additionally, one or more patch electrodes 56 can be used as reference electrodes 58, such that they operate in a single-ended mode. One patch electrode 56 in any pair of patch electrodes 56 can be used as the reference electrode 58 of that patch pair, forming a single-ended patch pair. One or more patch pairs can be configured to be independently single-ended. One or more patch pairs can share a patch as a single-ended reference, or reference patches of more than one pair of patches can be electrically connected.

[0071] Processing performed via the cardiac information console 20 allows axes to be transformed (e.g., rotated) from a first orientation (e.g., a non-physiological orientation based on the placement of electrode 56) to a second orientation. The second orientation can include a standard left-posterior-superior (LPS) anatomical orientation, where the "x" axis is oriented from the patient's right to left, the "y" axis from the patient's anterior to posterior, and the "z" axis from the patient's caudal to cranial side. The placement of patch electrodes 56 and the non-standard axes thus defined can be selected to provide improved spatial resolution than normal physiological orientations that result in the resulting axes, for example, due to preferred tissue characteristics between the non-standard oriented electrodes 56. For example, non-standard electrode placement can result in reduced negative impacts of low-impedance volumes in the lungs on the localization field. Furthermore, electrode placement can be selected to generate axes that traverse the patient's body along paths of equivalent or at least similar lengths. Axes of similar length will have more similar energy densities per unit distance within the body, producing more uniform spatial resolution along these axes. Transforming non-standard axes to standard orientations can provide the user with a more direct display environment. Once the desired rotation is achieved, each axis can be scaled as needed, i.e., made longer or shorter. Rotation and scaling are performed based on a comparison of the predetermined (e.g., expected or known) shape and relative dimensions of the electrode array 12 with measurements of the shape and relative dimensions of the electrode array in a coordinate system corresponding to the patch electrodes. For example, rotation and scaling can be performed to transform a relatively inaccurate (e.g., uncalibrated) representation into a more accurate one. Shaping and scaling the representation of the electrode array 12 can adjust, align, and / or otherwise improve the orientation and relative dimensions of the axes for more accurate positioning.

[0072] The electrical reference electrode 58 may be, or at least comprises, a patch electrode and / or an electrical reference catheter, which serves as a patient “analog ground” reference. The patch electrode 58 may be placed on the skin and may be used as a return current for defibrillation (i.e., providing a secondary purpose). The electrical reference catheter may include a monopolar reference electrode for enhancing common-mode suppression. The monopolar reference electrode or other electrodes on the reference catheter may be used to measure, track, correct, and / or calibrate physiological, mechanical, electrical, and / or computational artifacts in cardiac signals. In some embodiments, these artifacts may be due to respiratory, cardiac motion, and / or applied signal processing (e.g., filtering). Another form of electrical reference catheter may be an internal analog reference electrode, which may serve as a low-noise “analog ground” for all internal catheter electrodes. Each of these types of reference electrodes may be placed in relatively similar locations, such as in an internal vessel near the lower back (as a catheter) and / or on the lower back (as a patch). In some embodiments, system 100 includes a reference catheter 58, which includes a fixation mechanism (e.g., a user-activated fixation mechanism) that may be configured and arranged to reduce displacement (e.g., accidental or other unintentional movement) of one or more electrodes of the reference catheter 58. The fixation mechanism may include mechanisms selected from the group consisting of: a spiral expander; a spherical expander; a circumferential expander; an axially driven expander; a rotary-driven expander; and combinations of two or more of these.

[0073] exist Figure 1 The image depicts aspects of the receiver component of the cardiac information console 20. The cardiac information console 20 includes a defibrillation (DFIB) protection module 22 connected to a connector 20a, configured to receive cardiac information from the catheter 10. The DFIB protection module 22 is configured to have a precise clamping voltage and minimal capacitance. Functionally, the DFIB protection module 22 serves as a surge protector, configured to protect the circuitry of the console 20 during the application of high energy to a patient, such as during defibrillation (e.g., using a standard defibrillator).

[0074] The DFIB protection module 22 is coupled to three signal paths: a bioelectric potential (BIO) signal path 30, a location (LOC) signal path 40, and an ultrasound (US) signal path 60. Typically, the BIO signal path 30 filters out noise and retains the measured bioelectric potential data, and also enables the reading (e.g., successful recording) of the bioelectric potential signal during ablation, which is not the case in other systems. Typically, the LOC signal path 40 allows high voltage input while filtering out noise from the received location data. Typically, the US signal path 60 uses an ultrasound transducer 12b to acquire range data from the physical structure of the anatomical structure to generate a 2D or 3D digital model of the cardiac chamber HC, which can be stored in memory.

[0075] BIO signal path 30 includes an RF filter 31 coupled to DFIB protection module 22. In this embodiment, RF filter 31 operates as a low-pass filter with high input impedance. High input impedance is preferred in this embodiment because it minimizes voltage loss from a source (e.g., catheter 10), thereby better preserving the received signal, for example, during RF ablation. RF filter 31 is configured to allow biopotential signals from electrodes 12a on catheter 10 to pass through RF filter 31, for example, through frequencies less than 500 Hz, such as frequencies in the range of 0.5 Hz to 500 Hz. However, high frequencies, such as high-voltage signals used in RF ablation, are filtered out from biopotential signal path 30. In some embodiments, RF filter 31 may include a corner frequency between 10 kHz and 50 kHz.

[0076] BIO amplifier 32 may include a low-noise single-ended input amplifier that amplifies the RF filtered signal. BIO filter 33 (e.g., a low-pass filter) filters noise away from the amplified signal. BIO filter 33 may include a filter of approximately 3 kHz. In some embodiments, BIO filter 33 includes a filter of approximately 7.5 kHz, for example when system 100 is configured to adapt to cardiac pacing (e.g., to avoid significant signal loss and / or attenuation during cardiac pacing).

[0077] BIO filter 33 may include a differential amplifier stage for removing common-mode power line signals from biopotential data. This differential amplifier may implement a baseline recovery function that removes DC offset and / or low-frequency artifacts from the biopotential signal. In some embodiments, the baseline recovery function includes a programmable filter that may include one or more filter stages. In some embodiments, the filter may include a state-dependent filter. The characteristics of the state-dependent filter may be based on thresholds and / or other levels of parameters (e.g., voltage), and the filter rate may vary based on the filter state. Components of the baseline recovery function may include noise reduction techniques, such as pulse-width modulation and / or jittering of the baseline recovery voltage. The baseline recovery function may also determine the filter response of one or more stages through measurement, feedback, and / or identification. The baseline recovery function may also determine and / or distinguish signal portions representing the physiological signal morphology of artifacts from the filter response and computationally recover the original morphology or portions thereof. In some embodiments, the recovery of the original morphology may include subtracting the filter response directly and / or after additional signal processing of the filter response, such as by static, time-dependent, and / or spatially dependent weighting, multiplication, filtering, inversion, and combinations thereof. In some embodiments, the baseline recovery function may be implemented in BIO filter 33, BIO processor 36, or both.

[0078] The LOC signal path 40 includes a high-voltage buffer 41 coupled to the DFIB protection module 22. In this embodiment, the high-voltage buffer 41 is configured to accommodate relatively high voltages used in therapeutic techniques, such as RF ablation voltages. For example, the high-voltage buffer may have a power rail of ±100V. In some embodiments, each high-voltage buffer 41 has a high input impedance, such as 100 kΩ to 10 MΩ at the positioning frequency. In some embodiments, all high-voltage buffers 41 together are electrically equivalent in total parallel connection and also have a high input impedance, such as 100 kΩ to 10 MΩ at the positioning frequency. In some embodiments, the high-voltage buffer 41 has a bandwidth that maintains good performance over a high frequency range (e.g., frequencies between 100 kHz and 10 MHz, such as approximately 2 MHz). In some embodiments, the high-voltage buffer 41 does not include a passive RF filter input stage, for example, when the high-voltage buffer 41 has a power supply of ±100V. A high-frequency bandpass filter 42 may be coupled to the high-voltage buffer 41 and may have a passband frequency range of approximately 20 kHz to 80 kHz for positioning. In some embodiments, filter 42 has low noise with unity gain (e.g., a gain of 1 or about 1).

[0079] US signal path 60 includes a US isolation multiplexer MUX 61, a US transformer with Tx / Rx switches, a US transformer 62, a US generation and detection module 63, and a US signal processor 66. The US isolation MUX 61 is connected to the DFIB protection module 22 and is used to turn the US transducers 12b on / off, for example, in a predetermined sequence or mode. The US isolation MUX 61 may be a high input impedance switch group that, when on, isolates the US system and the remaining US signal path elements, decoupling the ground impedance (through the transducers and US signal path 60) from the inputs of the BIO path and LOC. The US isolation MUX 61 also multiplexes a transmit / receive circuit to one or more transducers 12b on the catheter 10. The US transformer 62 operates in both directions between the US isolation MUX 61 and the US generation and detection module 63. The US transformer 62 isolates the patient from the current generated by the US transmit and receive circuitry in module 63 during ultrasound transmission and reception at the US transducers 12b. US transformer 62 can be configured to selectively engage the transmitting and / or receiving electronics of module 63 based on the operating mode of transducer 12b, for example, by using a transmit / receive switch. That is, in transmit mode, module 63 receives a control signal from US processor 66 (within data processor 26), which activates US signal generation and connects the output of the Tx amplifier to US transformer 62. US transformer 62 couples the signal to US isolation MUX 61, which selectively activates US transducers 12b. In receive mode, US isolation MUX 61 receives reflected signals from one or more transducers 12b, which are then passed to US transformer 62. US transformer 62 couples the signal to the receiving electronics of US generation and detection module 63, which in turn transmits the reflected data signal to US processor 66 for processing and use by user interface system 27 and display 27a. In some embodiments, processor 66 commands MUX 61 and US transformer 62 to allow the transmission and reception of ultrasound to activate one or more associated transducers 12b, for example, in a predetermined order or mode. As an example, the US processor 66 may include detecting a single first reflection, detecting and identifying multiple reflections from multiple targets, determining velocity information from a Doppler method and / or from subsequent pulses, determining tissue density information from the amplitude, frequency, and / or phase characteristics of the reflected signal, and combinations of one or more of these.

[0080] Analog-to-digital converter (ADC) 24 is coupled to BIO filter 33 of BIO signal path 30 and high-frequency filter 42 of LOC signal path 40. ADC 24 receives separate sets of time-varying analog biopotential voltage signals, one for each electrode 12a. These biopotential signals have been differentially referenced to the unipolar electrode on a channel-by-channel basis for enhanced common-mode rejection, filtering, and gain calibration. Also received by the ADC via LOC signal path 40 are separate sets of time-varying analog positioning voltage signals for each axis of each patch electrode 56, which are output to ADC 24 as a collection of 48 positioning voltages measured simultaneously (in this embodiment) by electrode 12a. ADC 24 features high oversampling to allow noise shaping and filtering, for example, at an oversampling rate of approximately 625 kHz. In some embodiments, sampling is performed at or above the Nyquist frequency of system 100. ADC 24 is a multi-channel circuit that can combine BIO and LOC signals or separate them. In one embodiment, as a multi-channel circuit, the ADC 24 can be configured to accommodate 48 positioning electrodes 12a and 32 auxiliary electrodes (e.g., for ablation or other processes), for a total of 80 channels. In other embodiments, more or fewer channels may be provided. Figure 1 In this embodiment, for example, almost all the components of the cardiac information console 20 can be repeated for each channel (e.g., except for the UI system 27). For example, the cardiac information console 20 may include a separate ADC for each channel, or 80-channel ADCs. In this embodiment, signal information from the BIO signal path 30 and the LOC signal path 40 is input to and output from the respective channels of the ADC 24. The outputs from the channels of the ADC 24 are coupled to the BIO signal processing module 34 or the LOC signal processing module 44, which preprocess their respective signals for subsequent processing, as described below. In each case, the preprocessed signals are prepared for processing by their respective dedicated processors, as discussed below. In some embodiments, the BIO signal processing module 34 and the LOC signal processing module 44 may be implemented wholly or partially in firmware.

[0081] The biopotential signal processor module 34 can provide gain and offset adjustment and / or digital RF filtering, featuring a non-dispersive low-pass filter and an intermediate frequency (IF) band. The IF band can eliminate ablation and localization signals. The biopotential signal processor module 34 may also include a digital biopotential filter, which can optimize the output sampling rate.

[0082] Additionally, the biopotential signal processor module 34 may also include "pacing blanking," which is the blanking of information received during, for example, a physician "pacing" the heart within a time frame. Temporary cardiac pacing can be achieved, for example, by inserting or applying intracardiac, intraesophageal, or percutaneous pacing leads. The goal of temporary cardiac pacing is to interactively test and / or improve heart rhythm and / or hemodynamics. To achieve this, active and passive pacing triggering and input algorithm triggering determination can be performed, for example, by system 100. Algorithm triggering determination can use channel subsets, edge detection, and / or pulse width detection to determine whether pacing has occurred in the patient. Optionally, pacing blanking can be applied to all channels or a subset of channels, including channels where no detection has occurred.

[0083] Pacing silencing

[0084] Figures 17A to 17C The figure illustrates a non-limiting example of an embodiment for pacing blanking of one or more channels. One or more channels can be grouped for a variety of reasons, including but not limited to functional similarity (e.g., surface ECG), locational similarity (e.g., channels sharing a physical system board by system architecture), or convenience. In the figures, the boxes within the dashed boxes (in Figure 17 and labeled “N channels”) operate on individual channels, while the remaining boxes (… Figure 17A C) then operates on the channel group.

[0085] Ideally, the incoming data is corrected for channel-specific gain and offset. The input data can also be filtered or decimated at the input, for example, to reduce computational complexity and / or meet data rate requirements. The data stream is then split into two paths: a delayed path and a hold path, which converge at the output data switch.

[0086] The delay path is delayed by a duration similar to or equal to the processing time of the data window, which is processed to detect pacing pulses, for example, a delay of 0.1 to 100 ms, or 0.5 to 10 ms, or 4 ms.

[0087] The hold path includes a filter and a data latch. The hold filter is configured with a frequency response to exclude higher frequency components of the signal or power line frequency and variations in the low-frequency baseline, such as a bandpass filter with a passband range of 1 to 40 Hz. The hold filter may also include the duration of the averaged data to smooth any discontinuous transitions in the output signal of the output data switch, such as durations from 1 microsecond to 1 second, from 10 to 500 microseconds, or even 100 microseconds. The data latch of the hold path is configured to hold the value of its input for a specific duration, or until released by a control input command.

[0088] The output data switch is configured based on the blanking control logic to switch its output between the signal obtained from the delayed path and the signal obtained from the hold path. When the output data switch outputs the hold path signal, the signal is considered to be blanked. When the output data switch outputs the delayed path signal, the signal is considered not to be blanked.

[0089] The blanking control logic is the final stage of the pacing detection path. It uses input data to determine the presence or absence of pacing pulses within the current window of the input data. For each channel in the group containing input data, the channel is individually included or excluded in the subset used to detect pacing pulses. An enable signal for each channel is received from an external input and used to include or exclude each channel in the subset. The external input can be automatically determined by System 100 or manually configured by the user. For each channel included in the subset, a processing algorithm is used to process the signal. The processing algorithm can include, but is not limited to, one or more of the following: filtering, derivative, integration, downsampling, upsampling, absolute value, averaging, statistical calculations such as median, etc.

[0090] Each channel of the subset is aggregated into a summing node, whose operating algorithm combines information from all channels of the subset into a representative output. The algorithm used by the summing node may be, but is not limited to, one or more of the following: sum, average, difference, interleave, etc.

[0091] The representative output of the summation node is passed to the pacing pulse detection algorithm. The pacing pulse detection algorithm may include, but is not limited to, one or more of the following: slope detector, duration measurement, amplitude detector, threshold, derivative, absolute value, average value, filter, etc. The sequence of criteria used for pacing pulse detection can be arranged logically or as a decision tree based on the state of each criterion. One or more criteria can be enabled to allow pacing blanking, one or more criteria can be enabled to disable pacing blanking, or a combination of these. The result of the pacing pulse detection algorithm is passed to the blanking control logic to control the output of the output data switch.

[0092] The blanking control logic includes communication functions for transmitting the status of pacing pulse detection between channel groups. The output status reports to all other groups whether a pacing pulse is detected in the current channel group. The input status receives the status of one or more other channel groups. The blanking control logic can be configured to combine the received status with the result of pacing pulse detection in the blanking control logic to overturn the result of pacing pulse detection and / or ignore the received status.

[0093] Additionally, the biopotential signal processor module 34 may also include a dedicated filter that removes ultrasound signals and / or other unwanted signals, such as artifacts, from the biopotential data. In some embodiments, edge detection, threshold detection, and / or timing correlation may be used to perform such filtering.

[0094] The positioning signal processing module 44 can provide separate channel / frequency gain calibration, IQ demodulation with tuned demodulation phase, synchronous and continuous demodulation (without multiplexing), narrowband R-filtering and / or time filtering (e.g., interleaving, blanking, etc.), as described below. The positioning signal processing module may also include digital positioning filtering, which optimizes the output sampling rate and / or frequency response.

[0095] In this embodiment, the algorithmic calculations for BIO signal path 30, LOC signal path 40, and US signal path 60 are performed in the cardiac information console 20. These algorithmic calculations include, but are not limited to: processing multiple channels simultaneously, measuring propagation delays between channels, converting x, y, z data into a spatial distribution of electrode positions, including calculating and applying corrections to the set of positions, combining individual ultrasound distances with electrode positions to calculate detected endocardial surface points, and constructing a surface grid from the surface points. The number of channels processed by the cardiac information console 20 can be between 1 and 500, for example between 24 and 256, such as 48, 80, or 96 channels.

[0096] Data processor 26 may include one or more of various types of processing circuitry (e.g., microprocessors) and memory circuitry, executing computer instructions necessary for processing preprocessed signals from BIO signal processing module 34, positioning signal processing module 44, and US TX / RX MUX 61. Data processor 26 may be configured to perform calculations, as well as data storage and retrieval, which are necessary for the functionality of cardiac information processing system 100.

[0097] In this embodiment, the data processor 26 includes a biopotential (BIO) processor 36, a location (LOC) processor 46, and an ultrasound (US) processor 66. The biopotential processor 36 can perform processing on biopotentials, such as those recorded, measured, or sensed from the electronics 12a. The LOC processor 46 can perform processing on location signals. And the US processor 66 can perform image processing on, for example, US signals reflected from the transducer 12b.

[0098] The biopotential processor 36 can be configured to perform various calculations. For example, the BIO processor 36 may include an enhanced common-mode rejection filter, which may be bidirectional to minimize distortion, and may seed common-mode signals. The BIO processor 36 may also include an optimized ultrasound reflection filter and be configured for selectable bandwidth filtering. Processing steps of the data in signal path 60 may be performed by the biosignal processor 34 and / or the bioprocessor 36.

[0099] The positioning processor 46 can be configured to perform various calculations. As discussed in more detail below, the LOC processor 46 can electronically correct (calculate) axes based on the known shape of the electrode array 12, correct for scaling or skew of one or more axes based on the known shape of the electrode array 12, and perform "fitting" to align the measured electrode positions with known possible configurations, which can be optimized by one or more constraints (e.g., physical constraints, such as the distance between two electrodes 12a on a single spline, the distance between two electrodes 12a on two different splines, the maximum distance between two electrodes 12a, the minimum distance between two electrodes 12a, and / or the minimum and / or maximum curvature of the spline, etc.).

[0100] US processor 66 can be configured to perform various calculations associated with the generation of US signals via US transducer 12b and the processing of reflected US signals received by US transducer 12b. US processor 66 can be configured to interact with US signal path 60 to selectively send and receive US signals to and from US transducer 12b. US transducers 12b can each be set to either a transmit or receive mode under the control of US processor 66. US processor 66 can be configured to construct 2D and / or 3D images of the cardiac chamber (HC) in which electrode array 12 is disposed using reflected US signals received from US transducer 12b via US path 60.

[0101] The cardiac information console 20 may also include positioning drive circuitry, comprising a positioning signal generator 28 and a positioning drive current monitoring circuitry 29. The positioning drive circuitry provides a high-frequency positioning drive signal (e.g., 10 kHz to 1 MHz, or 10 kHz to 100 kHz). Positioning using these high-frequency drive signals reduces the influence of cellular responses, such as those from blood cell deformities, on positioning data, and / or allows for higher drive currents, for example, to achieve a better signal-to-noise ratio. The signal generator 28 generates a high-resolution digitally synthesized drive signal (e.g., a sine wave) with ultra-low phase noise timing. The drive current monitoring circuitry provides a high-voltage, wide-bandwidth current source, which is monitored to measure the impedance of the patient P.

[0102] The cardiac information console may also include at least one data storage device 25 for storing various types of recorded, measured, sensed and / or calculated information and data, as well as program code embodying the functions available from the cardiac information console 20.

[0103] The cardiac information console 20 may also include a user interface (UI) system 27 configured to output the results of localization, biopotential, and US processing. The UI system 27 may include at least one display 27a to present these results graphically in 2D, 3D, or a combination thereof. In some embodiments, the display 27a may include two simultaneous views of the 3D results, each having independently configurable view / camera characteristics such as view orientation, zoom level, pan position, and object characteristics such as color, transparency, brightness, and so on. The UI system 27 may include one or more user input components, such as a touchscreen, keyboard, and / or mouse.

[0104] Automatic Orientation

[0105] As described above, system 100 can use a combination of the following: voltage measured on electrodes 12a (placed within the internal volume of the heart, such as the left atrium or other chambers of the heart), the measurement location of these electrodes 12a, and an electronic model of the heart surface to generate an image. System 100 may include a positioning system for determining the location of electrodes 12a. Furthermore, signals from ultrasound transducers 12b can be combined with positioning information to estimate the heart surface (e.g., to create an electronic model of the heart surface). System 100 may include multiple sets (e.g., three sets) of patch electrodes 56 positioned nominally orthogonal to each other to generate three positioning axes: the X-axis (X1-X2), the Y-axis (Y1-Y2), and the Z-axis (Z1-Z2), as shown below. Figure 1 As shown. Each set of patch electrodes 56 includes pairs of skin electrodes placed at various locations on the body. The patch electrodes 56 shown in dashed lines are located on the patient's back. System 100 can be configured to provide a current (e.g., AC current) flowing through the distributed impedance between the pairs of patch electrodes 56, generating a variable voltage distribution in space (e.g., in the patient tissue between the patch electrodes 56). This voltage can vary over time based on the morphology of the waveform applied by system 100 (e.g., amplitude-modulated waveform, phase-modulated waveform, sweep waveform, and / or chirp). The resulting time-varying signals (in-phase and quadrature) are demodulated synchronously using a fixed phase relationship between the interrogation demodulation signal and the source applied to the patch electrodes 56. This demodulation results in a one-to-one correspondence between three sets of signal measurements generated by the patch (e.g., measured voltages) and their positions along three axes. This configuration allows for the spatial positioning of electrodes 12a.

[0106] The positioning coordinate system can be in an orientation different from standard (normal) physiological orientation; in other words, different from anterior-posterior, cranio-coccygeal, and left-right. In some embodiments, system 100 is configured to perform automatic orientation of the positioning coordinate system, which orients one or more positioning axes based on a desired orientation. This automatic orientation is achieved by comparing positioning voltages measured from electrodes having a known orientation relative to patch electrode 56 (e.g., ECG electrodes on the body and / or other electrodes of system 100) with information regarding the actual orientation (placed by the user) of patch electrode 56 on the patient's body. Alternatively or additionally, system 100 can be configured to perform automatic orientation, which rotates the positioning coordinate system to match standard physiological orientation. The ability to deterministically orient (e.g., via transfer function) the positioning coordinate system to match standard physiological orientation is useful, especially when using differentially driven positioning systems, where the measured voltages can vary with the positioning (position) of a reference patch or electrode on or inside the body.

[0107] System 100 can be configured to orient one or more positioning axes in a desired direction according to the following steps:

[0108] In the first step, the direction of voltage variation between two or more electrodes placed in a known orientation relative to the positioning axis can be used to establish the desired orientation of the positioning axis, as defined by each pair of patch electrodes 56.

[0109] In some embodiments, ECG leads are placed at fixed locations on the body relative to the heart. This placement provides a set of electrodes with a known orientation relative to the heart chambers and can be used to establish the orientation of the positioning axis. In some embodiments, the positioning X-axis spans the left-right direction of the body, and two ECG leads (e.g., V5 and V6) are positioned on the left side of the heart chambers. The voltages from these ECG leads are then compared with the voltages from electrodes (e.g., electrode 12a) inside or near the heart volume to establish the orientation of the X-axis.

[0110] Reference electrode 58 can be positioned on the patient's lower back, near one of the patch electrodes 56, thereby generating the Y-axis (e.g., near...). Figure 1 (Y1 shown). This placement of the reference electrode 58 positions the origin of the Y-axis on the lower back. The voltage of the reference electrode 58 can be set to zero volts (e.g., the reference electrode 58 is grounded); and the orientation of this positioning axis can be established by comparing it with the voltage measured from electrodes placed near the heart (e.g., electrode 12a and / or other electrodes).

[0111] In the second step, system 100 can use a coordinate system that satisfies the right-hand rule. This use provides additional constraints on the orientation of the axes that establish the positioning coordinate system. The current flowing between the patch electrodes 56 indicates the orientation of the patch electrodes 56, which in turn establishes the orientation of the positioning axes. The direction of current flow can also be established using the direction of voltage variation (e.g., gradient). In a triaxial system with current injected in a nominally orthogonal direction, the principal components of the voltage distribution also indicate or approximate the orientation of the patch. In some embodiments, the right-hand rule condition is used in conjunction with the estimated directions of current from the three axes in the heart chambers to fix the orientation of the Z-axis.

[0112] Once the desired orientation of the positioning axis (e.g., the axis defined by the pair of patch electrodes 56) is established, the system 100 can rotate the positioning coordinate system to a physiological orientation based on the knowledge of the nominal position of the patch electrodes 56. In some embodiments, the YZ plane is rotated approximately 45° to align with the physiological orientation (e.g., the natural physiological coordinate system).

[0113] Higher speed dipole density algorithm

[0114] In order to obtain the surface distribution of cardiac information (e.g., surface charge density and / or dipole density), system 100 can map the information measured from electrode 12a onto the endocardial surface of the atrium. This mapping can be performed as an inverse problem, and the number of measurement points (approximately 48 points from 48 electrodes 12a on array 12) is much smaller than the number of points that the desired solution (e.g., a solution including more than 3000 points on the atrial surface) is expected to be based on. The inverse problem is ill-posed and requires special methods to solve. In some embodiments, system 100 uses Tikhonov regularization [1] to solve the inverse problem according to the following objective function:

[0115]

[0116] in It is a data fitting term. is the regularization term, λ is the regularization parameter, s is the source on the surface, φ is the measured potential, and A is an m×n matrix, where m << n.

[0117] Related to Singular Value Decomposition (SVD), the singular value decomposition of matrix A is

[0118] A=U∑V (2)

[0119] Where U∈R m×m V∈R n×n It is a singular vector matrix, ∑∈R m×n It is a diagonal matrix with diagonal elements being decreasing singular values. SVD can be used to compute the pseudo-inverse of the least-squares solution of a linear system.

[0120] With respect to minimum norm estimation (MNE), system 100 can be configured to use SVD of matrix A, such that the regularization of the inverse problem, constrained by Equation 1, will be:

[0121]

[0122] Where V m =V(:,1:m),∑ m =∑(:,1:m) contains the first m columns from V and ∑.

[0123] In the early design, System 100 implemented the MNE of Equation 3, which typically took about 3000 milliseconds (3 seconds).

[0124] Due to the following EP-specific improvements to the solution method for the inverse problem, a continuous mapping of cardiac electrical activity can be achieved, as described later in this paper.

[0125] In some embodiments, system 100 is configured to operate more efficiently (e.g., to compute solutions to the inverse problem in near real-time or real-time, referred to herein as "real-time"), for example by using truncated SVD computation (e.g., instead of the full SVD computation described above) and / or CPU-optimized memory allocation algorithms.

[0126] System 100 may include an inverse algorithm, which includes C++ code containing the function CalculateTikhonovInverseDouble, such as Figure 1 As shown. SVD calculations are performed using the Intel Math Kernel C++ Library (MKL), a high-performance and fast library, especially suitable for Intel-based CPUs. Using MKL, matrix A is calculated for each time frame of the biopotential measurement on System 100. m×n All singular values. Later in the algorithm, the number of singular values ​​can be truncated to find a solution to Equation 3. Below is the interface to the C++ function that computes the solution to Equation 3:

[0127]

[0128] In its early design, System 100 implemented methods for performing SVD calculations, such as... Figure 1-1 As shown in (a). In step 152, a complete SVD is performed on the matrix:

[0129]

[0130] In step 154, the results of the performed calculations are combined and truncated according to the following equations:

[0131]

[0132] In step 156, the transpose is calculated, expressed as:

[0133]

[0134] In step 158, again, another complete SVD calculation is performed on the results from step 156 according to the following equation:

[0135]

[0136] Then, in step 160, the transpose is found, which is represented as:

[0137]

[0138] The truncated Tikhonov is calculated from the result of step 160.

[0139] Figure 1-1 Left side (i.e.) Figure 1-1 The implementation shown in (a) is computationally expensive and has room for performance improvement. Since Equation 3 only requires m singular values, it is not necessary to compute all singular values ​​of A, and the truncated SVD of A can be directly found in a single step 164 (using MKL, such as...). Figure 1-1 (as shown in (b)). Therefore, Figure 1-1 (b) method is better than Figure 1-1 The method in (a) is significantly more effective, for example, with lower computational intensity.

[0140] Figure 1-1 The implementation of (b) eliminates the need for several mathematical calculations in the previous implementation, further enhancing processor performance. This improvement includes secure elimination of transpose and additional SVD calculations. Figure 1-1 (a) Step 158 of the flowchart. In other words, using the new implementation, it is not necessary to perform the transpose and SVD calculations twice. In summary, Figure 1-1 The flowchart shown in (a) can be reduced to a single step of directly calculating singular values, as follows: Figure 1-1 As shown in (b).

[0141] The following code snippet represents the execution of the discussion above. Figure 1-1 (b) Step 164:

[0142] info = LAPACKE_dgesvd(

[0143] LAPACK_ROW_MAJOR,

[0144] 'S', 'S',

[0145] n, n,

[0146] inputMatrix->GetMatrix(),

[0147] Ida, S, U, Idu, V, Idvt, superb);

[0148] In this improved SVD implementation described above, only the singular values ​​required by Equation 3 are directly calculated.

[0149] As described above, the early design implemented the MNE of Equation 3, which resulted in approximately 3000 milliseconds in the inverse solution computation time. In some embodiments, system 100 includes an algorithm with improved memory allocation, such as... Figure 4 As shown. In experiments, it was found that memory allocation accounted for 70% of the total function time in the previous algorithm, while SVD accounted for approximately 30%. Due to the size of the matrix used for SVD computation, the high cost of memory allocation can be eliminated by using the ACMMatrix2D class to compute efficiently and serially (instead of using the GPU). For continuous mappings (e.g., changes in surface charge information and / or dipole density shown in real time), system 100 can use MKL memory allocation (i.e., mkl_malloc, which is optimized for Intel CPUs), such as... Figure 5 As shown.

[0150] In some embodiments, system 100 includes the improvements described above and achieves approximately 40 times faster performance with the same accuracy (e.g., the computation time for the inverse solution will be less than 100 milliseconds, such as approximately 70 milliseconds).

[0151] Below is a code example that uses the ACMMatrix2D C++ class to provide improved memory allocation for SVD calculation of A = U∑V:

[0152] if (U != NULL) delete U;

[0153] U = new ACMMatrix2D <double>();

[0154] U->SetSize(OnMainMemory,m,m,true);

[0155] if(V!=NULL)delete V;

[0156] V=new ACMMatrix2D <double>();

[0157] V->SetSize(OnMainMemory,n,n,true);

[0158] if(S!=NULL)delete S;

[0159] S=new ACMMatrix2D <double>();

[0160] S->SetSize(OnMainMemory, 1, minDim, true);

[0161] The following code provides an improved method for efficient memory allocation in SVD computation:

[0162] void ＊ mkl_malloc(size_t alloc_size, int alignment);

[0163] As an example, the following references can provide information related to the methods described above for mathematically solving such problems:

[0164] 1. ANTikhonov and VYArsenin, "Solutions to ill-posed problems", Mathematical Computation, Vol. 32, pp. 1320-1322, 1978.

[0165] 2. Intel, "Reference Manual for Core Mathematical Functions Library", 2007.

[0166] Map using multiple basket positions

[0167] Some embodiments use the positions of a single electrode array 12 (also referred to herein as "baskets") to compute the transfer matrix in the inverse mapping, which can limit the length of the mapping segment and the accuracy of the mapping. In other embodiments, each basket position can be used to compute the transfer matrix in the inverse mapping. Using each basket position for the transfer matrix can incur significant computational costs, slowing down the mapping workflow—achieving only minimal accuracy improvements for a subset of these positions. In another embodiment, a subset of positions where substantial accuracy improvements can be achieved can be algorithmically determined—for example, based on the relative changes in the basket positions.

[0168] For multiple basket positions, monitor the current position of each basket relative to its previous reference position, and if a basket changes position beyond a desired threshold, update the calculated basket position based on its current position. Figure 9 As shown. In one embodiment, the reference position of the basket can be its position in a previous time sample. In another embodiment, the reference position can be its position at a previous interval that is dynamically updated as the position changes beyond a threshold. In yet another embodiment, the reference position can be its position at a specific point in time that remains static throughout the computation. This algorithm provides mapping accuracy while saving processing power, which is associated with faster computation. In a 1 to 5-second recording, there may be 3 to 4 mm (or more) of uncertainty related to the basket's position (which doctors typically try to keep still during recording). In one embodiment, for an input mapping segment, the algorithm uses a position (e.g., a first position) as a reference position and then checks for changes in the position (e.g., to determine if the reference position should be adjusted). For example, if a 1 mm position change is detected, a position update can be triggered for the computation. This process is repeated until the end of the mapping segment. The electrocardiogram (EGM) and time are aligned with the basket position to compute an accurate mapping in the inverse kinematics. In another embodiment, the first position can be used as an initial reference position. The reference position is updated when a position change that triggers a position update in the computation is detected. In another embodiment, the reference position can be continuously updated. For example, when a new position is acquired and its position is compared with the reference position, the reference position can be updated to the newly acquired position for the next measurement.

[0169] Non-functional electrode detection

[0170] System 100 uses voltages measured at the locations of electrodes 12a (e.g., at least 24 electrodes and / or approximately 48 electrodes in a basket or other array configuration) to record voltage sets for calculating dipole density and / or surface charge density on the endocardial surface. Unreliable voltage measurements can harmfully affect any calculations involving positioning and / or biopotential recording, leading to poor positioning and / or abnormal estimations of cardiac information. Therefore, system 100 may include automated procedures for identifying any electrodes 12a ("non-functional" electrodes) suspected of having abnormal voltage measurements.

[0171] System 100 may include the algorithm described below: Algorithm NFE, for identifying nonfunctional electrodes 12a. The NFE algorithm employs three different strategies to identify the measured localization potential, which may not accurately reflect the true electrode environment at the time of measurement. This procedure can be performed during anatomical reconstruction, such that each frame of the anatomical recording (e.g., a video frame or other anatomical image, such as a frame representing a time increment of approximately 25 to 100 milliseconds) is associated with a list of electrodes 12a identified as nonfunctional during the data collection associated with that frame. Subsequent computations involving localization or biopotential recordings can access the stored list of nonfunctional electrodes 12a associated with each frame, thereby eliminating unreliable recordings from critical computations.

[0172] In some embodiments, system 100 includes identifying non-functional electrodes for an NFE algorithm that operates (to a minimum) during data recording made for the creation and / or re-creation (here, "creation") of a digital heart model. The NFE algorithm can be configured to process continuous data and data interleaved with ultrasound data. The NFE algorithm can be configured to evaluate the measured voltage before it is correlated with spatial coordinates. The NFE algorithm can be configured to generate a list of non-functional electrodes 12a that differs from each frame containing a list of electrodes 12a from which unreliable potentials are measured.

[0173] Now for reference Figure 3 The details of method 300 are shown, which includes a series of steps of an NFE algorithm for detecting nonfunctional electrode 12a.

[0174] In the first step 302, an anatomical file including I / Q data, such as an anatomical file of at least one cardiac chamber, is loaded. In the second step 304, system 100 calculates I / Q amplitude and phase from the I / Q data. In the third step 306, the I / Q amplitude data is converted into potential data. In step 308, nodes from anatomical structures with anomalous phases are identified. In step 310, each axis is scaled by the maximum voltage range based on the potential data. In step 312, nodes with anomalous radii relative to the mean are identified. In step 314, the covariance of the voltage data is calculated. In step 316, the data is rotated to the basis of the covariance eigenvector. In step 318, a best-fit triaxial ellipsoid is calculated for the rotated data. In step 320, outliers relative to the surface of the ellipsoid are identified. This allows for the identification of non-existent faults (NFEs).

[0175] More specifically, the NFE algorithm includes various inputs, such as I Data and Q data Each is an integer array of size N multiplied by the number of channels (numCH) x 3. The first dimension includes the number of samples, the second dimension includes the number of channels, and the third dimension includes the frequencies of the three positioning axes. I / Q data (complex voltages measured at electrode 12a) can be derived from recorded files or real-time streaming data. Another input could be... gain_rawdataToVolts It includes floating-point numbers, which limit the ADC count-to-voltage conversion gain of the system 100 hardware. gain_rawdataToVolt A typical value could be 0.52074e-6. The other input could be an array of integers. refnodes _ aux The array contents specify the index of auxiliary catheter electrodes that can be used for motion correction.

[0176] The nonfunctional electrode NFE algorithm may include one or more outputs. The outputs of the NFE algorithm may be a frame-specific array including a list of nonfunctional electrodes 12a from which voltage measurements are unreliable. In some embodiments, the user is notified of the identified list of nonfunctional electrodes. In some embodiments, the user can manually select any electrode 12a to be removed, regardless of whether the selected electrode is part of the identified list of nonfunctional electrodes. In some embodiments, the system 100 allows the user to accept the removal of identified nonfunctional electrodes; this is an opt-in condition. In some embodiments, the system 100 allows the user to accept the removal of a subset of identified nonfunctional electrodes; this is a partial opt-in condition, where the subset is determined by user input. In some embodiments, the system 100 recommends a subset of identified nonfunctional electrodes 12a for removal, where this subset is determined by an algorithm that identifies electrodes 12a identified as nonfunctional based on a set of time conditions (e.g., functional duty cycle and / or continuous duration). In some embodiments, the system 100 automatically removes the list of nonfunctional electrodes 12a from display, from subsequent calculations, or a combination of both. In some embodiments, system 100 automatically removes the list of non-functional electrodes 12a from calculations of the device's position, orientation, and / or shape fit.

[0177] In some embodiments, unreliable electrode voltage measurements in each frame are identified through one, two, or three different procedures, which are described separately below.

[0178] Identifying outliers based on I / Q phase anomalies

[0179] Each electrode 12a can be associated with the amplitude and phase of the positioning potential involving the three positioning axes (i.e., the angles corresponding to the real and imaginary components measured by the positioning voltage in the complex plane). The positioning phase can be set (e.g., adjusted) at the start of a procedure (e.g., a clinical procedure or part of a clinical procedure) such that all functional electrodes 12a exhibit a phase close to a set value (e.g., somewhere away from 0 and / or pi). The mean and standard deviation can be calculated from the phase values ​​of each electrode 12a. Electrodes 12a exhibiting a phase above a threshold (e.g., differing from the mean by more than a constant multiplied by the standard deviation) are identified as nonfunctional and added to the list of nonfunctional electrodes associated with the current frame.

[0180] Identifying outliers based on potential relative to the mean.

[0181] In some embodiments, the voltage of a subset of electrodes 12a exhibiting phase consistent with a selected value is adjusted to a zero mean, and a maximum range is calculated along each positioning axis. The potential of electrode 12a on each positioning axis is scaled to the maximum range on that axis, and the radius of the scaled point (relative to the vanishing mean) is calculated. Any point exhibiting a radius greater than a threshold (e.g., a specific constant multiplied by the average radius) is identified as an anomaly and loaded into the list of non-functional electrodes associated with the current frame.

[0182] Identify points that deviate from the best-fit ellipsoid

[0183] The voltage of the remaining electrode 12a was adjusted back to zero mean. The covariance of the data array was calculated using the following formula:

[0184] C ij =P ik P kj

[0185] The repeated exponent means that the number of remaining electrodes 12a is summed after eliminating the non-functional electrodes 12a identified in any one or two of the steps described above.

[0186] The eigenvectors of the covariance matrix are calculated through singular value decomposition:

[0187] C = USV t

[0188] The columns of the square array U are the desired eigenvectors.

[0189] The basis for transforming voltage data into feature vectors:

[0190] P′ k =V kl P l

[0191] Where P l It is the 3-D potential associated with a specific electrode 12a.

[0192] The voltage is scaled relative to an orthogonal reference, and the data is fitted to a generalized triaxial ellipsoid with a semi-major axis aligned with the basis vectors:

[0193]

[0194] a, b, and c are the semi-major axes along the three coordinate directions, respectively. Fitting can be performed either by using the explicit LLS method or by calculating the pseudo-inverse using singular value decomposition (SVD).

[0195] A semi-major axis is defined, and three potential sets associated with a specific electrode k can each be assigned an error:

[0196]

[0197] The average error is calculated, and any point exhibiting an error greater than a threshold (e.g., a specific constant multiplied by the average error) is identified as a non-functional electrode and loaded into the list of non-functional electrodes associated with the current frame.

[0198] The cascade of nonfunctional electrodes 12a identified by one or more (e.g., all) of the above three steps includes a list of nonfunctional electrodes 12a associated with the current frame that exhibit voltages considered unreliable for calculations including on-site estimation (e.g. for positioning) or cardiac information (e.g., dipole density and / or surface charge density).

[0199] Recursive recognition

[0200] After identifying non-functional or "bad" electrodes (regarding the positioning potential) and calculating the local field estimation / scaling matrix, locating all electrodes 12a (including those identified as non-functional) and aligning them with a standard basket (e.g., the geometry of an array of electrodes 12a) allows for subsequent evaluation of the separation of the electrodes from their respective "standard" locations. Electrodes 12a identified as non-functional can still be located close to their respective "standard" locations. Such "good" electrodes 12a can be removed from the list of non-functional electrodes 12a, thus providing the maximum number of electrodes available for calculations such as field estimation and / or auxiliary electrode positioning.

[0201] Figure 10 This is a flowchart illustrating an example of the steps for reclassifying nonfunctional or "bad" electrodes for calculating cardiac activity. Step 1 includes recording positioning and biopotential data using a multi-electrode catheter (e.g., catheter 10). Step 2 filters the data for nonfunctional or "bad" electrodes using the criteria discussed above for determining good / bad (first criterion). In some embodiments, the first criterion is biased towards identifying electrodes as bad. Step 3 then filters the identified nonfunctional or "bad" electrodes using a second criterion to determine whether the "bad" electrodes can be reclassified as "good" again. In some embodiments, the second criterion is biased towards identifying electrodes as bad. If some nonfunctional or "bad" electrodes are reclassified as "good," the data from the final group of "good" electrodes is used to calculate cardiac activity data. In one embodiment, step 3, filtering the identified non-functional electrodes to reclassify them as "good," can be performed by: a) replacing one or more test subsets of the identified non-functional electrodes into the "good" classification; b) determining the contribution of the electrodes in the test subsets to the error in the basket positions; and c) repeating steps (a) and (b) for all permutations / combinations (or subsets) of the electrodes included in the test subsets. This repetition can be performed, for example, recursively. The final set of electrodes returning to the "good" classification can be determined by a comparison method, for example, the largest test subset with a total error contribution below a threshold, or the test subset with the lowest error contribution among all test subsets. The contribution of the electrodes in the test subsets to the error in the basket positions can be determined, for example, by quantifying the basket position variations with and without electrodes from the test subsets as translations, rotations, general affine transformations, or the sum, mean, and / or RMS of paired distances between matching electrodes in two basket positions (e.g., an array of electrode 12a positions). The basket location for determining the error can be the original, local electrode location or the electrode location derived from geometric fitting (e.g., SVD fitting).

[0202] ADC card on / off

[0203] In some embodiments, system 100 includes a subsystem that shuts down ADC 24 under certain conditions, such as to improve system efficiency or to reset the data stream to clear data storage, such as buffers, filters, or memories. For example, system 100 may be configured to shut down ADC 24 to reduce load, thereby reducing the data throughput to processor 26 or other components of system 100. In some embodiments, system 100 is configured to shut down ADC 24 when calculating inverse kinematics to determine dipole density and / or surface charge density from voltage data. In some embodiments, system 100 is configured to shut down ADC 24 whenever exiting an acquisition mode, such as a mode that collects (e.g., actively collects) data from electrode 12a and / or ultrasonic transducer 12b. System 100 may be configured to automatically, based on user input, or a combination of both, shut down ADC 24. In some embodiments, system 100 is configured to shut down ADC 24 and automatically turn it back on after a period of time as a reset function. Reset may be time-based automatic (periodic), input-based automatic (triggered), or user-based manual. In some embodiments, processes that are in parallel or in series with the ADC 24 (e.g., filtering, buffering, blanking algorithms, interleaving, etc.) can be configured to be disabled in an appropriate order to minimize artifacts caused by transient changes in the ADC 24, such as being enabled from a disabled state or disabled from an enabled state. In some embodiments, after the ADC 24 has been reset, the processes that were disabled to minimize artifacts can automatically return to their operating state before the ADC 24 was reset.

[0204] V-Wave Removal (VWR)

[0205] System 100 can be configured to at least partially remove the ventricular far-field signal (V wave) that can be applied to the atrial signal of interest (e.g., to assess arrhythmias such as atrial flutter and / or atrial fibrillation). For example, system 100 may include an algorithm (VWR algorithm) configured to identify and remove components of the intracardiac electrogram (EGM) resulting from ventricular excitation. The VWR algorithm may employ a process of analyzing recordings from electrode 12a and user input. The ventricular far-field signal (V wave) can be significantly subtracted from the EGM while atrial activity is well preserved.

[0206] After at least some signal filtering, the VWR algorithm can operate on the recorded data (to a minimum). The VWR algorithm can assume that the user has selected the appropriate segment corresponding to the V wave on the EGM. The VWR algorithm can assume that the V wave is consistent throughout the recording.

[0207] refer to Figure 4 The diagram shows a block diagram that details the entire sequence of steps in method 400 required for the V-wave removal (VWR) algorithm.

[0208] The VWR algorithm may include one or more inputs. Inputs to the VWR algorithm may include the EGM of the data record, such as an EGM including atrial activity and ventricular components. Inputs to the VWR algorithm may include start and end sample numbers, which may represent a ventricular segment, and may limit the user-selected ventricular example for the V wave in the EGM.

[0209] The VWR algorithm can include one or more outputs. The output of the VWR algorithm can include the estimated V component in the input EGM. The output of the VWR algorithm can include the EGM minus the V wave.

[0210] The V-wave removal (VWR) algorithm can be configured as a filter or pseudo-filter (here, "filter") and / or include a filter. The start and end sample points can be input parameters for this filter. The filter can be used to remove the ventricular component from the EGM.

[0211] In some embodiments, the VWR algorithm includes three steps: 1) automatically identifying the V-band, steps 402 and 404; 2) creating a V-band template (e.g., a channel-specific V-band template), steps 406 and 408; and 3) subtracting the V-band from the EGM and smoothing the boundary, step 410.

[0212] For example, refer to Figure 4 In the first step 402, the system (e.g., processor 10) automatically selects the peak value as the reference time point and segmentation interval. In step 404, the system calculates the cross-correlation value ("cc value") between the sample V segments and the selected V segments; V segments with cc < 0.9 are rejected. In step 406, all V segments are averaged to generate a V template. In step 408, the system aligns the V template and adjusts the time shift according to the time delay. Furthermore, in step 410, the system smooths the boundaries of the V template to create smooth transactions and avoid sudden jumps.

[0213] In some template-based V-wave removal algorithms for intracardiac electrograms (EGM), large atrial signals with amplitudes comparable to ventricular signals can lead to the detection of some atrial signals as ventricular segments, which can result in the removal or alteration of the atrial signals intended for mapping. By using simultaneous recording of the EGM and ECG, ventricular timing from the surface ECG channel can be identified through peak detection. Since the atrial to ventricular signal amplitudes are relatively small (typically visible in the ECG compared to the intracardiac ECM), the ECG channel can be preferred as an auxiliary reference for ventricular timing selection. The peaks detected from the ECG channel are then used as timing references to preserve timing from the EGM (e.g., timing difference <= 50 ms). Timing associated with ventricular activity is then preserved, and atrial activity-related timing is subsequently removed from further ventricular template generation and subtraction.

[0214] Figure 11A and 11B Two examples of V-wave removal (VWR) are shown. In each figure, a top trace is collected from an ECG lead showing a large atrial signal, a middle trace is an electrocardiogram of System 100, and a bottom trace is an electrocardiogram with the V-wave removed. In these examples, the V-wave and subsequent high-pass filter artifacts are selectively removed.

[0215] V-band identification

[0216] To remove the V-wave signal from the EGM, the initial step is to identify segments that include the V-wave component. Since the V-wave measured on the biopotential channels (e.g., voltage measurements on electrode 12a) has a consistent waveform and amplitude, the BCT voltage (an estimated voltage at the geometric center of the electrode 12a array) can be calculated by averaging all channels at each time point. The peak value in the BCT can then be used as a reference time point for the V-wave.

[0217] Then, user-specific V-wave samples can be used to locate V-bands. Peaks in the V-wave samples can be aligned with a reference time point. The maximum cross-correlation between the reference point segment and the V-wave samples can be calculated. The initial alignment can then be adjusted based on the time delay generated by the maximum cross-correlation. The aligned segment in the EGM can then be marked as a V-band.

[0218] Create channel-specific V-wave templates

[0219] Once the V-bands are identified in the 48 channels, channel-specific V-band templates can be created by aligning them with the V-bands within that channel and taking their average value. Figure 4-1 Curve "A" in the figure shows an example of a V-wave template for a single channel.

[0220] Figure 4-1 Curve "B" in the diagram shows an example of the EGM after removing the V wave. The ventricular far-field signal is reduced to 80%. Atrial activity and baseline are preserved.

[0221] Subtract V-wave and smooth the boundary

[0222] Once a channel-specific V-band template is created, the data from the V-band is subtracted from the V-band in the EGM. However, "jumps" can occur in the signal from the V-band at the beginning and end points of the V-band template. See [link to example of this phenomenon] for an example. Figure 4-2 The Hanning window can be used to smooth edges, for example... Figure 4-3 As shown in the image.

[0223] V-wave disappearance or zeroing

[0224] In other embodiments, blanking or zeroing of the V wave in the EGM can be used instead of V wave subtraction, removing any ventricular contribution from the signal while accepting the atrial signal. In some embodiments, a signal segment containing V wave data can be selected manually (e.g., using a "caliper" on the screen) and / or using a template-based detection method (e.g., system 100 automatically detects the V wave using a signal template). The segment can be blanked or "zeroed" by setting all values ​​in the segment to a set value (e.g., zero) and / or by interpolation across the segment duration (e.g., linear interpolation between data values ​​just before or after the selected time period). Smoothing at or near the boundaries of the segment's beginning and end samples can be used to minimize transition artifacts from interactions with other filters.

[0225] V-wave blanking or zeroing with reduced baseline offset

[0226] In some embodiments, blanking or zeroing of the V wave in the EGM is performed using the original signal (e.g., before applying a filter such as a high-pass filter (HPF)). As mentioned above, zeroing of the V wave performed after applying HPF can result in morphological artifacts and / or atrial baseline shifts due to the inclusion of large and unbalanced signals (e.g., V waves) in the HPF processing.

[0227] Alternatively, V-wave zeroing performed prior to HPF application can prevent atrial baseline shift by removing large and unbalanced signals (e.g., V-waves) before HPF application and processing. In some embodiments, the interpolation of V-wave zeroing prior to HPF application matches the slope and curvature of any low-frequency components (e.g., respiration or drift).

[0228] Activation detection of propagated historical images

[0229] Figure 12 This is an embodiment of a display of cardiac activation data. In the right-hand view, the heart is shown from a first-person perspective, showing the various activation bands. The outermost band is the band of activated nodes. Subsequent bands represent the most recently activated states. Activation states relate to time indicated by a sliding window overlaid on the EGM, which has a horizontal time axis. The width of the sliding window reflects a 50ms window width, referred to as the "propagation history." In the left-hand view, different perspectives of the same heart are shown. This view shows that the activation waves have propagated around the heart. Various user interaction devices can be provided in the display for manipulating the 3D image of the heart, such as rotating the image, adding labels to the image, etc.

[0230] In various embodiments, the cardiac information dynamic display system may include video display control tools that allow users to play, replay, rewind, fast forward, pause, and / or control the playback speed of a dynamic activation wave propagating on the heart.

[0231] Automatic annotation of activation times is crucial for generating propagation history images. However, noise in the virtual EGM can introduce artifacts into the propagation history images, which not only cause visual noise but also lead to less accurate activation detection.

[0232] Median filter used for propagating historical images

[0233] A median filter can be applied to the propagation history image. For each node in the cardiac anatomy, the propagation history value is updated to the median of the original propagation history values ​​of the node and its neighbors. These neighbors are limited to nodes within a specific distance of the node being updated. As this distance increases, more noise is removed, but less local propagation history detail is preserved. For further filtering, the median filter can be applied multiple times to the propagation history image.

[0234] A fundamental principle behind using median filters for post-processing propagation history is that the propagation history image can be considered a series of images. Noise in the propagation history most closely resembles "salt and pepper noise," a term often referenced in image processing. Traditional low-pass filters (blurring filters) struggle to remove salt and pepper noise without introducing other artifacts (such as distorted boundaries) in the image. Median filters can remove salt and pepper noise while preserving sharp boundaries (see...). Figure 13 ).

[0235] Figure 13 Examples of median filters applied to propagation history images are shown. A) The original frame of the propagation history image. B) A median filter has been applied to the propagation history in A), resulting in a significant reduction in artifacts compared to A). C) A second median filter has been applied to the image in B), resulting in even fewer artifacts compared to A) and C).

[0236] False positive activation removal

[0237] Automatic annotation of activation times is crucial for generating propagation history images. However, noise in the virtual EGM can introduce spurious activations, which may lead to incorrect interpretation of the propagation history images.

[0238] In some embodiments, an algorithm is used to detect noise from true activations by examining the amplitude of an electrogram within a time window surrounding the potential activation. If the electrogram exceeds an amplitude threshold, the potential activation is referred to as a true activation. This size of the time window surrounding the potential activation serves not only as a strict amplitude threshold but also as a minimum slope threshold. Larger time windows allow for the exposure of lower-slope activations (e.g., those found in slowly depolarizing tissue), while smaller time windows around activations will reveal regions of faster depolarization.

[0239] Figure 14 An exemplary implementation of false-positive activation annotation removal is shown. A) Initially detected activations are shown in green. B) An example of removing false positives from activations is circled in A). Within a time window, the amplitude of the electrogram is examined. Since this activation does not meet the amplitude threshold, it is removed in C). C) Activations automatically detected after false-positive removal. Low-amplitude / slope activations from A) are removed.

[0240] Registration process

[0241] In some embodiments, system 100 includes a registration processing subsystem (e.g., an algorithm) to associate electric field measurements (voltage measurements) with location and maintain coordinate registration between the located intracardiac device and the reconstructed cardiac surface displayed on the UI throughout the procedure, adapting to intentional and unintentional changing conditions, such as power loss, system hardware or software restarts, changes in the location source output, and / or changes in the distribution of the in vivo location field. The location system of system 100 plays a crucial role in acquiring and analyzing cardiac information, such as guiding ablation catheters and / or providing accurate location information for mapping and other sensing catheters. The final step in location is calculating the coordinates of electrodes or other catheter components from their location voltage readings. Inputs and outputs are defined as follows:

[0242] Catheter assembly coordinates (X, Y, Z) = F(catheter positioning voltage (Vx, Vy, Vz)),

[0243] Where F is a function that maps voltage space to coordinate space.

[0244] The mapping function F (registration) between voltage space and coordinate space is limited by the following factors:

[0245] 1. Positioning systems, such as those described in co-pending U.S. Provisional Application No. 62,161,213, filed May 13, 2015, entitled "Positioning System and Method for Acquiring and Analyzing Cardiac Information," the entire contents of which are incorporated herein by reference for all purposes.

[0246] 2. The patient.

[0247] Because the function F varies under different conditions (e.g., different reference electrode 56 positions, different patient anatomy, etc.), it cannot be a static part of system 100. Therefore, the function F must be determined dynamically and appropriately stored for patient treatment and potential post-procedural analysis.

[0248] In some embodiments, the voltage-to-location mapping function F is configured as follows.

[0249] 1. Function F contains information that converts voltage into coordinates, including but not limited to:

[0250] a. The conduit in use (e.g., type or mode), the list of system input channels, and the list of excluded channels (e.g., channels with non-functional electrodes 12a as described herein).

[0251] b. Scaling transformation between voltage and coordinates

[0252] c. Optionally, a position or coordinate reference measured or calculated from the system input channel group (e.g., electrodes, patches, etc.), including a list of system input channels, a list of excluded channels, and the original or initial voltage.

[0253] 2. Database storage of registration data

[0254] a. A master copy of the registration data is saved together with the corresponding anatomical 3D model.

[0255] b. A supplementary copy of the registration data is saved along with each record, including:

[0256] i. Content from the main copy

[0257] ii. Updates after obtaining the primary copy, such as the different exclusion channels and location reference information for the conduits.

[0258] 3. Register the search path

[0259] When registration data is required for calculations, whether for post-analysis or real-time guidance of ablation catheters, cardiac information algorithms (e.g., dipole density and / or surface charge density algorithms as described herein) will search for the correct registration by following these paths:

[0260] auxiliary dungeons

[0261] If no corresponding secondary copy is found, the primary copy will be searched.

[0262] If the primary replica is not found, a warning message is displayed to the user and the default configuration is used.

[0263] 4. Alignment of independent registration configurations

[0264] When the independent primary configuration and independent secondary configuration do not match, such as when an inappropriate anatomical structure is assigned to the analysis record, the cardiac information algorithm will perform registration correction to align the secondary configuration with the primary configuration, or vice versa, such as ensuring that the resulting coordinate systems of the two configurations are unified and can be used in the same coordinate space.

[0265] 5. Detection and correction of registration changes

[0266] Under changing conditions, such as power loss, system hardware or software restarts, changes in positioning source output, and / or changes in the distribution of the positioning field within the body, the registration processing subsystem can detect such changes, notify the user, automatically reload previous settings, automatically perform corrections to a state consistent with the configuration before the change, or any combination thereof. In some embodiments, automatic correction may include measurements from electrodes and / or sensors within the body (e.g., intracardiac electrodes) and / or on or near the body surface (e.g., positioning source patches or ECG electrodes).

[0267] Location or coordinate reference

[0268] As a component of the registration, the measured input set can be used to stabilize, reduce, or eliminate undesirable, time-varying variations in the positioning system. These can include periodic variations due to respiration or cardiac motion, discrete variations due to the introduction, removal, and / or manipulation of devices within the positioning field (especially with low input impedance), and / or slow variations associated with physiological or clinical phenomena such as patient sweating, saline infusion, adhesive layers of skin patches or electrodes, and interactions between electrogels and / or hydrogels and the skin. The measured inputs can be:

[0269] 1. Directly used to cancel time-varying changes using subtraction to eliminate common-mode signal components.

[0270] 2. Used in conjunction with weights (or linear models), where the weights are determined by the following:

[0271] a. Analytical models, such as those derived from field theory or principal component analysis.

[0272] b. Empirical derivation of the data set for training dynamic algorithms.

[0273] c. Filtering

[0274] Combinations of da, b and / or c

[0275] 3. Used in conjunction with models based on higher-order functions, which may be nonlinear models.

[0276] 4. Used in conjunction with geometric transformations (e.g., affine transformations) or geometric minimization fitting methods.

[0277] Input can consist entirely of electrodes inside the body, or it can consist entirely of surface patches and electrodes, or a combination of these.

[0278] When the measured input deviates from the quantitative model or the expected output, the threshold can be used to trigger a user message or warning on the UI.

[0279] In some embodiments, system 100 includes a physical location reference comprising a device (e.g., a device including at least one electrode) positioned at a fixed location within a coordinate system near a "region of interest" (e.g., an area to compensate for drift and / or other motion artifacts, such as near and / or within a heart chamber). The device may be positioned within a portion of an anatomical structure such that it is impossible for it to move relative to the region of interest, e.g., when positioned within a coronary sinus. The device's position is then assumed to be stationary relative to the region of interest (e.g., by the positioning processor 46), and any movement of the device is determined to be caused by motion artifacts. These motion artifacts are removed from location data (e.g., positioning data) collected relative to all other devices within the region of interest. Motion artifacts can be caused by: breathing; cardiac activity; sweat absorption by the positioning patch; patch adhesion degradation during the procedure; saline infusion during the procedure; changes in bodily conduction; introduction of one or more conductive and / or insulating materials into the patient (e.g., an additional catheter); introduction of one or more devices connected to a low-impedance ground source into the patient; and combinations of one or more of these.

[0280] Alternatively or additionally, system 100 may include a "virtual coordinate" reference, also referred to herein as a "virtual position" reference. System 100 may be configured to "learn" (or be "trained") to provide a virtual position reference, as described below. Physiological and / or environmental influences (e.g., respiratory effects or other motion effects that produce artifacts) on the positioning system can be measured from the body surface using "artifact calibration signals" recorded by external electrodes located on the patient (e.g., surface patches X1, X2, Y1, Y2, Z1, Z2 and / or 58 as described herein, and / or any other similar patches placed on the skin). The artifact calibration signals may be recorded as a drive signal is provided to one or more external electrodes (e.g., a drive signal provided by positioning signal generator 28) and / or as the drive signal is between one or more external electrodes. System 100 may be trained (as part of a calibration or "training" procedure) to recognize these influences such that a physical position reference is not required within the patient's body, and instead, the virtual position reference is calculated using signals measured by the external electrodes. To train system 100, a physical training device (e.g., device 10, the ablation catheter of system 100, and / or any catheter containing electrodes) is positioned within the patient near the region of interest and kept stationary, and the physical training device is positioned using system 100 as described herein. The positioning signal typically associated with the motion of the physical training device (e.g., by positioning processor 46) is assumed to be caused by one or more artifacts of the positioning system (i.e., because the physical training device remains stationary). During this positioning, external electrodes simultaneously record artifact calibration signals. The recorded positioning signals are then correlated with the simultaneously recorded artifact calibration signals, and mathematical processing is performed to establish a representational relationship between the positioning system and the artifact calibration signals recorded by the external electrodes. The representational relationship may include transformations to be applied to all devices within the positioning field (e.g., transformation to a virtual position reference). Mathematical processing may include: fitting methods; a series of weighted coefficients; linear systems; machine learning; adaptive fitting and / or filtering methods; principal components; filtering (e.g., low-pass and / or high-pass filtering); other mathematical methods; and / or combinations of one or more of these.

[0281] After training, the physical training device can be removed (or reused, e.g., if the catheter 10 of system 100 is used as a physical training device), and system 100 can use transformations (e.g., with improved accuracy by eliminating motion artifacts and other physiological and / or environmental influences) to effectively locate devices within the field. In some embodiments, training can be repeated manually and / or automatically to confirm and / or update the transformations. For example, system 100 can periodically perform diagnostics (e.g., using one or more algorithms) to determine whether a user is moving any electrodes or devices located within the patient's body, and if not, training of system 100 as described above (e.g., in the background) can begin. Once training is complete, the newly generated mathematical transformations can be compared with previous transformations. If the transformations differ, they can be updated using the newly calculated transformations. In some embodiments, buffers are used to store historical data, thereby processing data over longer periods (e.g., transformations can be adjusted over time based on cumulative training).

[0282] Manual registration

[0283] Modeling and automatic compensation for stabilizing, reducing, or eliminating undesirable time-varying variations in the positioning system can result in residual offsets / errors with static or constant components. Any constant component of the overall compensation or residual can be applied by assigning a constant value (i.e., a constant vector) to each coordinate axis. This constant vector can be determined through user interaction with the user interface and system display for manual calibration. The system display can show one or more views of the positioned device in relation to a positionally stable or fixed object in the coordinate system (e.g., the surface of a heart chamber, a model, and / or a set of points).

[0284] User interaction with an input device (such as a mouse, keyboard, and / or touchscreen) can activate a mode that allows displacement of the positioned device relative to a fixed object. This mode can be activated via keys, buttons, and / or other commands integrated into the workflow, or it can be a separate window or interface with different entry and exit modes.

[0285] The displacement of the positioning device relative to the fixed object can be performed on recorded data or applied to real-time streaming data, so that the positioning device continues to move under the control of its application or the motion applied within the body, adding the displacement applied by the user to the calculated coordinate position, which is based on the mapping function F, as described above, or any other registration component.

[0286] In some embodiments, the displacement is applied only to the display's visual engine and must be confirmed by the user before being applied to the mapping function F or any other positioning engine calculation. The application of the mapping function F may be temporary, or it may be permanent, or until it is modified.

[0287] User interaction with the system can be visually or graphically guided by a set of surface information data acquired at or near the interaction time. The acquired surface information data set can be provided by sensors or electronics in contact with the surface, and / or it can be provided by non-contact sensors or sensor arrays. Non-contact sensors or sensor arrays can be optical sensors, ultrasonic transducers, and / or other ranging or imaging modalities. Simultaneous visual or graphical display of the acquired surface information can be used to help determine the displacement to be applied by aligning two surfaces or objects. In some embodiments, the imaging modality can be ultrasonic surface reflection. In one embodiment, a set of ultrasonic vectors with a length equal to the extent to the surface can be displayed. The endpoints of the vectors can then be aligned with a fixed object (e.g., a model of a heart chamber, a surface, and / or a set of points). In another embodiment, a new surface can be partially or fully reconstructed and can be used to determine the displacement to be applied by aligning the new surface with a fixed object or a heart chamber surface, model, and / or set of points.

[0288] Automatic or assisted imaging-mediated registration

[0289] Alignment of the acquired surface information data set with a fixed object (e.g., a heart chamber or surface) can be performed based on an algorithm. The algorithm may include minimizing distance, angle, and / or other cost functions or geometric quantities. Minimization can be performed on points, vectors, lines, planes, and / or other geometries. The algorithm may be analytical and / or iterative. The algorithm may also include the calculation of affine transformations to modify the position, shape, scaling, and / or other geometric properties of the acquired surface information data set to improve alignment. The modified geometric properties may include all affine transformations or be limited to one or two. In some iterative embodiments, criteria (exit conditions) are used to terminate the iteration. Criteria may include error, minimum error, maximum error, mean square error, root mean square error, aggregate error, total error, and / or other conditions of the cost function. In some embodiments, the minimization technique may use an iterative nearest-point algorithm.

[0290] In one embodiment, the alignment algorithm uses a newly acquired set of points from a multi-directional non-contact measurement method (e.g., ultrasonic ranging), which can be transformed to optimally match a fixed reference object (e.g., a cardiac chamber surface, model, or set of points). The transformation can be limited to translation. The newly acquired points can be directly compared to the initially acquired set of points. Alternatively, the newly acquired points can be used to calculate a representative surface and compared with the initially acquired set of points, or the initially acquired set of points can be used to calculate a representative surface and compared with the newly acquired set of points, or both the newly acquired and initially acquired points can be used to calculate two representative surfaces to be compared.

[0291] Use field estimation for positioning

[0292] In some embodiments, system 100 includes an improved field estimation (IFE) algorithm. Locating a device or its components within a body cavity or chamber using electrical impedance techniques requires an accurate and precise understanding of the electric field at all points within the chamber. Device localization can be performed using methods, in part or in whole, combined with the methods described in International PCT Patent Application PCT / US2016 / 032420, filed May 13, 2016, entitled "Location System and Method for Acquiring and Analyzing Cardiac Information," which is incorporated herein by reference. Given the possibility of field non-uniformity, the (relative) direction and magnitude of the field in space between any pair of points must be known, thereby obtaining the (relative) position through the integration of field gradients. In some embodiments, system 100 includes enhanced field identification for catheter localization as described below.

[0293] The first step in positioning the catheter can be identifying the electric field within the cardiac chamber of interest. The catheter 10 can be configured, together with the console 20, to detect multiple AC electric fields (e.g., three fields oriented along three mutually perpendicular or nearly perpendicular directions and distinguished by frequency) at multiple spatially distinct locations (e.g., the positions of electrodes 12a, such as 48 electrodes 12a) on a spherical surface (e.g., a spherical surface approximately 1 inch in diameter). After obtaining voltage samples at each electrode 12a, at each known frequency, the direction, amplitude, and gradient of the electric field relative to the known geometry of the array 12 can be calculated using least squares and / or other error minimization techniques. As the array 12 moves within the body cavity or chamber, electric field records are obtained for each electrode 12a relative to the geometry of the array 12. Due to the unknown orientation of the array 12, the priori position and orientation of the array 12 relative to an arbitrarily chosen origin can be obtained by integrating the spatially varying field along the path connecting each electrode 12a to the origin.

[0294] As an example, the following references can provide information related to the methods described above used for mathematically solving such problems:

[0295] 1. Kabsch, Wolfgang, (1976) "Solution to the optimal rotation of two sets of vectors", Acta Crystallographica 32: 922 (revised in Kabsch, Wolfgang (1978) "Discussion on the optimal rotation of two sets of vectors", Acta Crystallographica, "A34", 827-828)

[0296] In some embodiments, the IFE algorithm processes data recorded during anatomical / geometric acquisition. In some embodiments, the IFE algorithm is capable of processing continuous data and data interleaved with ultrasound data. In some embodiments, the IFE algorithm utilizes a position reference to correct for common-mode error sources, such as those occurring during respiration (mechanical / impedance) and / or cardiac motion. The position reference may be a separate catheter that moves synchronously (common-mode) with array 12 when not manipulated.

[0297] In some embodiments, the IFE algorithm does not require an auxiliary internal device as a position reference. In some embodiments, the position calculation of array 12 is performed without the need for a fixed-shape array 12.

[0298] refer to Figure 5 The diagram illustrates a block diagram detailing an embodiment of the IFE algorithm. The diagram includes sections related to ultrasound and / or surface reconstruction, and describes flow control, field estimation, and assisted localization.

[0299] The IFE algorithm includes one or more inputs. For example, the I and Q data are integer arrays of size N x number of channels (numCH) x 3. The first dimension N is the number of samples, the second dimension is the number of communications, and the third dimension corresponds to the three positioning axis frequencies. The I / Q data can originate from recorded files and / or real-time streaming data. The IFE algorithm can include inputs. gain_ rawdataToVolts This is a floating-point number that limits the ADC 24 count-to-voltage conversion gain of the system hardware. In some embodiments, gain_rawdataToVolts This includes values ​​of approximately 0.52074e-6. The IFE algorithm can include input... stdbskdata This is an array of size 3, consisting of the number of electrodes 12a (e.g., 48). These values ​​will be the standard (e.g., elastic bias) XYZ positions of the electrodes 12a in the array 12. These values ​​are used to perform fitting and field calculations. The IFE algorithm may also include input... refnodes_aux This is an integer array. refnodes_aux The content specifies the index of the auxiliary catheter electrode that can be used for motion correction.

[0300] The IFE algorithm includes one or more outputs. The output of the IFE algorithm is the coordinates of all connected electrodes in system 100 and the field estimate (in the form of a scaling matrix) associated with the centroid position of each array 12. The coordinates are given as an Nx3 floating-point array, while the scaling matrix is ​​a 9-element array associated with each system of the anatomical / geometric record.

[0301] In some embodiments, the IFE algorithm is configured as follows.

[0302] Field identification and basket center positioning

[0303] Identify the origin system. This identification serves as the primary system for recording anatomical / geometric data. After transformation from the local system, all moment centers and auxiliary position vectors relate to this reference system. The basis vectors in the origin system are defined by standard (e.g., offset) position measurements of the electrodes 12a on array 12.

[0304] Field identification (determination of field strength / orientation) at each center of the radius:

[0305] Calculate the spatial field gradient (i.e., the rate of change of position as a function of potential) at each moment center location. Each coordinate dimension (x, y, z) in the local reference frame will produce a 3x3 array of scaling factors as a function of each source potential. This matrix is ​​the voltage-to-position transformation in a standard array 12 reference frame appended to a specific moment center:

[0306]

[0307] The scaling matrix is ​​calculated by representing the position vectors between 12 pairs of electrodes in the standard array 12 reference frame as a linear function of the differences between the three measured voltages between the same two electrodes:

[0308]

[0309] The subscripts "i" and "j" represent the positions of electrodes 12a on array 12. Each index operates on all electrodes 12a (e.g., 48 electrodes), except for any electrodes identified as unusable by the non-functional electrode algorithm as described herein. This is performed for each possible pair of electrodes 12a:

[0310]

[0311] Where "del" denotes the positional differential with respect to each of the three source potentials, and dx denotes the vector of x-coordinate differences (in this case). Ideally, we obtain NumChan for the three derivative terms on each coordinate axis. 2 The equation (where NumChan equals the number of electrodes 12a) is used. The resulting overdetermined system is solved in a least-squares sense to calculate nine "average" scaling factors defined by the known positions and potentials of the electronic devices 12a in three dimensions. Singular value decomposition A = USV is then used. t And the derivative along each coordinate axis is given by the following formula:

[0312]

[0313]

[0314]

[0315] Note that matrix A is identical on all three coordinate axes, so singular value decomposition only needs to be performed once. A robust scaling factor is obtained unless a large number of co-located electrodes are ignored (e.g., if all electrodes from 2 or 3 adjacent splines of array 12 are eliminated).

[0316] Matrix multiplication of the voltage values ​​using the scaling matrix yields the position in physical space relative to the origin of the local reference frame:

[0317]

[0318] The columns of the 9-element scaling matrix are identified as vectors obtained by moving along the gradient of the corresponding source voltage (e.g., the first column is a vector tangent to the gradient of the X source voltage). The field lines are not limited to being uniform or orthogonal, and a complete description of the field will require knowledge of the scaling matrix at every point in space.

[0319] System 100 can be configured to perform the following steps.

[0320] The positions of the array 12 electrodes 12a associated with each moment center in the local coordinates are reconstructed using a local scaling matrix. The local coordinates are again defined by the positions of the electrodes 12a in the standard array 12 (e.g., elastic bias positions). Note that the orientation of each local reference frame is unknown without knowing the electric field vector relative to some external reference frame.

[0321] Identify the path to the origin system in voltage space. Move along this path and find the center of magnitude (in voltage space) of the path closest to the identified origin. This set of centers of magnitude (including the target and the origin) constitutes an integration of the actual piecewise continuous paths to the origin. (See...) Figure 6 )

[0322] The distance between adjacent centers on the integration path is calculated using the scaling factor associated with the center closest to the origin (e.g., center 1 of segment 1). Figure 6 (Sections 1 to 4 in the text). Each distance is a position vector represented in a reference frame closer to the center of the moment (e.g., ...). Figure 6 (As shown). The position of the target moment center in the origin system is the vector sum of the distances between adjacent moment centers on the integration path, but these distances must first be represented in a consistent reference system (i.e., the origin system). Therefore, each distance interval along the integration path must be transformed into the origin system.

[0323] To determine the coordinate transformation between each moment center and the origin, system 100 constructs a series of transformations between each pair of adjacent moment centers along the integration path. The product of this set of transformations along the integration path is the desired transformation of the origin.

[0324] refer to Figure 7 In order to transform the farthest position vector (i.e. the separation between the target system and system 1) back to the origin system, we first need to find the transformation between system 1 and system 2.

[0325] The transformation is obtained by comparing arrays 12 in the two systems, which are reconstructed using the scaling matrix of the closer system (system 2 in this example). System 100 reconstructs array 12 in the more distant system (e.g., system 1) using the scaling matrix associated with the closer system (system 2). System 100 then uses the Kabsch algorithm to align the two arrays 12 (i.e., the array 12 associated with system 2, and the array 12 in system 1 reconstructed using the scaling matrix in system 2). The resulting transformation transfers coordinates from the more distant system (1) to the closer system (2). System 100 uses a series of transformations obtained by aligning adjacent arrays 12 as described above to transform each distance interval (between the centers of squares) back to the origin system. System 100 then transforms the next farthest position vector (segment 2, see again) for the next farthest position vector. Figure 6 Repeat the procedure. The vector sum of the transformed distance intervals is the position of the target moment center in the origin reference frame. Repeat this process for each moment center using the uniquely identified origin path and the associated coordinate transformation set obtained by applying the Kabsch algorithm between adjacent moment centers.

[0326] Therefore, for each target frame, system 100 obtains a transformation set that, when applied to the position vector obtained using a local scaling factor associated with the target moment center, restores the position vector in the point reference frame:

[0327]

[0328] For example, [T] 1，2 ] represents an orthogonal (i.e., length-preserving) transformation from reference frame 1 to reference frame 2.

[0329] Refer again Figure 6 The net effect is to transform position vector 1 into the origin coordinate system. Combined with similarly transformed position vectors 2 to 4, this (vector) sum represents the position of the target system in the origin coordinate system.

[0330] The orientation of the origin relative to global coordinates is obtained by comparing the electric field vector (columns of the scaled matrix in the origin system) with the basis vectors in the origin system, which is, for example, defined by the potential excitation axis (i.e., the reference frame of reference electrode 58). If and e j It is the unit basis vector formed by the coordinate axes defined by the standard array 12 and the electric field vector in the origin system. The components of this transformation are given by the following equation:

[0331]

[0332] Where "i" and "j" are vector indices (independent of subscripts 1, 2, etc., in the transformation used to identify the center of the moment along the integrated path for locating the target center of the moment). Note that both vectors must be given in a Euclidean reference frame to constrain the inner product.

[0333] Although the electric field vectors in the origin frame are known (as columns of the scaling matrix), they can actually be approximated by the eigenvectors of the covariance of the array 12 voltages in the origin frame. These vectors are known in the origin reference frame, and the above transformation applies. The advantage of this method is that the eigenvectors of the covariance matrix are orthogonal, and no subsequent orthogonalization is required.

[0334] It also obtains additional transformations into, for example, the AP reference frame.

[0335] After identifying the field (e.g., a 9-element scaling matrix where each moment center position is known) and defining the transformation of the position vector from each moment center (target system) to the origin system, the position of any electrode near the identified field is obtained as follows (see...). Figure 7 ).

[0336] Identify a subset of the moments that are “close” to the target electrode in the voltage space. A larger number of moments does not necessarily guarantee more accurate localization, for example, because increased separation in the voltage space may lead to a decrease in the fidelity of the electric field identification (i.e., the field at the moment center may not represent the field at the electrode location).

[0337] For each identified center of the moment, the potential offset from the average basket position (i.e., the position of array 12) is multiplied by a local scaling matrix to obtain the position vector relative to that center of the moment in local coordinates.

[0338] As mentioned above (reference) Figure 6 The transformation of the electrode position from the target system to the origin system is provided by a series of transformations, enhanced in each case by transforming from the target system to system 1. The electrode position is then the average of the sum of the moment center position and the local offset (represented in the reference origin system).

[0339]

[0340] The average of the above quantities at all identified centers of moment is the estimated position of the electrode in the origin system.

[0341] Note that the non-uniformity of the field in the volume of interest can increase the error in the location estimation because the distance from the volume where the field is well identified increases.

[0342] It can also be transformed into patch or AP system.

[0343] The in vivo application of the aforementioned enhanced positioning technology has demonstrated an accuracy of approximately 1 mm in a static environment (i.e., an environment where the position can remain stable).

[0344] Field identification with improved resolution and spatial dimensions

[0345] In the aforementioned methods for field identification and basket center-of-matrix localization, the estimation of the localization field utilizes a set of spatiotemporally correlated potentials measured from a device with a known geometry. The field estimate at the center-of-matrix of a known geometry represents the stability and readiness of the internal volume of that known geometry. The localization field can vary at or near impedance variation boundaries, such as near the endocardial surface or ostia connecting anatomical structures, such as veins, valves, and appendages. The field estimate at the center-of-matrix of a known geometry can limit the spatial dimension for achieving complete accuracy of the field estimate. The localization field in the region between the farthest location of the center-of-matrix of a known geometry and the navigable volume of the device with a known geometry can be identified by directly sampling this region using a subset of electrodes, rather than by measuring the full representative volume of the device.

[0346] To estimate local field variations in a region at high resolution, field estimation can also be achieved by decoupling the spatial and temporal correlations of field sampling data from the measuring device. In some embodiments, geometric constraints of known geometries (e.g., those of the electrode array 12), such as the full distance matrix between all arrangements of electrode pairs of the device, can be relaxed to facilitate finer sampling resolution.

[0347] In one embodiment, such estimation of local field variation can be achieved using only the known distances between adjacent electrode pairs. For a single electrode pair, the field variation is sampled only in the direction of the distance vector between the electrodes, and the uncertainty or error increases with the angle between the sampled vector and the direction of field variation. Direction dependence can be reduced by eliminating the temporal correlation of the sampled data and aggregating field sampling data from multiple electrode pairs (and thus multiple sampled vectors) traversing a finite volume space over time. The finite volume space can be constructed by dividing the coordinate system into three-dimensional voxel sets. Over a period of time, the sampled localization field potentials of discrete electrode pairs can be “binned” (e.g., grouped as “boxes”) into voxel sets. In some embodiments, simplified general localization methods can be used to determine the bins, such as those described above for field identification and basket center localization. The size of the bins can be varied to enclose a statistically significant set of potential samples from the electrode pairs. The set of sampled vectors within the bins defines the estimation of the local field. This estimation can be determined by applying SVD to a set of equations that correlate the spatial field gradient to the spatial and potential magnitudes and directions of the sampled vectors. Figure 18A An example is shown.

[0348] In another embodiment, sampled potentials from any individual electrode traversing a finite volume of space within the positioning field can be aggregated over time. Again, the finite volume of space can be constructed by dividing the coordinate system into groups of three-dimensional voxels. Over a period of time, the sampled positioning field potentials of any individual electrode can be binned into groups of voxels. In some embodiments, simplified general positioning methods can be used to determine these bins, such as those described above for field identification and basket center positioning. The size of the bins can be varied to enclose statistically significant groups of individual potential samples. The potential difference between adjacent or nearby bins, and their relative direction and distance, define an estimate of the local field. This estimate can be determined by applying SVD to a set of overdetermined equations that correlate the spatial field gradient to the location and potential difference of adjacent or nearby bins. Figure 18B An example is shown.

[0349] The measured potential is localized, for example, from a navigation aid conduit, by applying a known gradient value to each voxel.

[0350] Positioning and shape modeling of auxiliary devices

[0351] Intracardiac field estimation improves the accuracy of localization of assistive devices (such as lasso catheters or coronary sinus mapping catheters) by eliminating variations in the localization field across the chambers. The method allows variations in the field across space to be incorporated with less distortion into the localization of catheters placed near the ventricular wall and in veins / aprons / valve structures. Field variations are captured by electronics 12a (during anatomical reconstruction) and processed to integrate the field variations.

[0352] In some embodiments, the electrode location can be obtained from an auxiliary catheter, for example... Figure 15 The nonlinear, nominally circular "lasso"-shaped conduit shown can be modeled in various ways, for example:

[0353] 1. Cubic splines are independently fitted to the located electrode data along each of the three spatial coordinate axes.

[0354] 2. Higher-order (e.g., fourth or fifth order) polynomials are independently fitted to the located electrode data along each of the three spatial coordinate axes.

[0355] 3. A priori unknown radius and oriented circle are fitted to the positioning electrode data, and the resulting circle can be modified to allow helical deformation relative to the rotational symmetry axis.

[0356] In some embodiments, the calculation method for shape modeling may be a combination of methods. In some embodiments, the shape modeling method may be changed based on measurements of the positioned electrode position group or user-defined changes, for example, when the variable-shape conduit is configured as a circular loop, a method suitable for shape modeling a loop is used; when the conduit is configured as a curve, a method suitable for shape modeling a curve is used; and when the conduit is configured as a line, a method suitable for shape modeling a line is used.

[0357] In some embodiments, shape modeling uses the physical specifications of the auxiliary conduit, such as electrode spacing, electrode radius, electrode length, radius of curvature, etc. In some embodiments, shape modeling is calculated using electrode positions located on a subset of available physical electrodes. In some embodiments, electrode positions can be excluded from shape modeling calculations as "invalid" based on geometric and / or temporal criteria.

[0358] refer to Figure 16 This illustrates a method for selecting points (e.g., 2D or 3D coordinates of locations) for shape modeling of a catheter. In step 1610, electrical signals (e.g., through one or more components of system 100) are recorded in relation to the locations of multiple electrodes on the catheter located within the patient's body (e.g., within the chambers of the patient's heart). These locations are stored as a set of points. In some embodiments, the catheter includes a variable-shape auxiliary catheter comprising multiple electrodes along a portion of its length. The catheter may include multiple electrodes, such as a catheter comprising at least 8, 12, or 20 electrodes. In step 1620, the center of gravity of the point set is calculated. In step 1630, each point in the point set is compared to the calculated center of gravity and / or its neighboring points. In some embodiments, the distance between each point and the center of gravity is compared to a threshold, such as a threshold equal to twice the expected radius of the shape being modeled. In some embodiments, the distance between each point and its neighboring points is compared to the same threshold or a different threshold. In step 1640, if the distance comparison exceeds the threshold, the point is determined to be invalid and removed from the point set (step 1645a). In some embodiments, a point is determined to be invalid if at least one, at least two, or all of the comparisons performed exceed a comparison threshold, or if at least one comparison exceeds the threshold.

[0359] After comparing all points in the point set, it is determined whether there are enough points remaining in the set to compute the shape model (step 1650). In some embodiments, the minimum number of points (e.g., electrodes) required to compute shape modeling is based on the order of the polynomial used for the computation. In some embodiments, there must be more than 8 points remaining, for example, more than 10 or 12 points. In step 1660, if the required minimum number of points is not met, the system may display the raw location data (e.g., overlaid on a 3D or 2D display of the heart chambers) and no shape modeling computation is performed. In step 1670, if the minimum number of points is met, the system may perform the shape modeling computation and display the best-fit model of the catheter. In some embodiments, the display output of the shape modeling shows a graphical representation of all electrodes on the physical catheter, regardless of whether they are connected to and positioned by the system 100. In some embodiments, electrodes connected to and positioned by the system 100 are graphically displayed differently from electrodes not connected to and positioned by the system 100, for example, by differences in color, size, and / or shape.

[0360] Exercise correction

[0361] Accurate catheter positioning and motion tracking in 3D electrophysiological mapping systems are essential for arrhythmia diagnosis and ablation treatment strategies. The system conceived in this invention offers high accuracy and is highly responsive to motion. In noisy environments, atrial fibrillation can generate high-frequency motion, visually perceived as catheter "jittering" in the procedure. This "jittering" motion can distract the physician and lead to unnecessary fatigue due to visual strain.

[0362] In some embodiments, the algorithm of System 10 filters the positioning signal to eliminate jitter motion while maintaining appropriate responsiveness of the duct motion. To filter out high-frequency noise or motion, it may be necessary to reduce the bandwidth of the IQ filter. However, this can introduce additional delay and increase response time. Optimal filter delay and frequency bandwidth are critical. A CIC (Cascaded Integral Comb) filter is the only option to achieve this. Compared to IIR, mean, and median filters, CIC filters perform better in terms of delay and similarly reduce jitter motion. Furthermore, it operates quickly (e.g., efficiently) in embedded (e.g., FPGA) and PC applications. The filter itself requires only addition and subtraction. Multiplication is only needed when scaling the output (if required) to make the filter gain equal to 1.

[0363] The characteristics of a CIC filter are determined by three parameters. Different combinations of these parameters provide varying bandwidths and noise attenuation while maintaining low computational complexity. These three parameters are: deferential delay (DD), number of segments (Nsect), and number of downsampled samples (Nds).

[0364] Scaling algorithm

[0365] System 100 can use inverse kinematics to calculate cardiac information, such as the surface charge density distribution and / or charge density distribution on the surfaces of cardiac chambers (e.g., the surfaces of atrial chambers). The location of electrode 12a within the chamber (e.g., the atrium) and the geometry of the chamber surface (e.g., the atrial surface) are two essential inputs for formulating the inverse kinematics. Since array 12 can be located at different positions within the chamber, the distance from the chamber surface to the nearest electrode 12a can vary significantly. This variation affects the resolution of the inverse kinematics. For example, the magnitude of the reconstructed charge density tends to be more dominant in regions spatially closer to electrode 12a than in regions spatially farther from electrode 12a.

[0366] System 100 may include the scaling algorithm described below, which uses a weighted minimum norm estimation method to mitigate the distance effect on the forward matrix (pre-scaling); then, a uniformly distributed charge source is used to calculate a post-scaling value for adjusting the output of the inverse solution accordingly (post-scaling). This technique significantly improves the localization accuracy of the focal source using the inverse solution.

[0367] After appropriate signal filtering, the scaling algorithm can operate on the recorded data (to a minimum). In some embodiments, the algorithm also performs V-wave subtraction, as described above. In these embodiments, the algorithm may assume that the user has selected an appropriate segment on the EGM corresponding to the V-wave, and / or the algorithm assumes that the V-wave is consistent throughout the recording.

[0368] refer to Figure 8 The block diagram illustrates an embodiment of a series of steps in the scaling algorithm, including pre-scaling and post-scaling.

[0369] The scaling algorithm of system 100 may include one or more inputs. In some embodiments, the inputs to the scaling algorithm are the same as or similar to the charge density algorithm of system 100, for example, when the scaling algorithm includes inputs relating to the location of array 12 electrodes 12a. The inputs may be vertices and / or triangles defining the surface geometry of the chamber (e.g., the surface geometry of the atrium). The inputs may include EGM measured from array 12 electrodes 12a.

[0370] The scaling algorithm of System 100 can provide one or more outputs, such as outputs including surface charge density, dipole density and / or other cardiac information related to the chamber surface.

[0371] In some embodiments, the scaling algorithm comprises two main components. First, a weighted minimum norm estimate is used to improve the Tikhonov regularization method; then, a uniform source distribution is used to compute post-scaling values ​​for adjusting cardiac information (e.g., charge density) from the first step.

[0372] Positive problem (described by the charge density on the surface of the atria)

[0373] The linear relationship between charge density S and the electric potential φ measured on the conduit is described by the following equation:

[0374] φ=As (4)

[0375] Where A is the measurement matrix, which encodes the relationship between the charge density on the atrial surface and the potential on each electrode 12a, which involves not only the distance from electrode 12a to the atrial surface but also the atrial geometry. Given the known source s and the position of electrode 12a, the voltage measured on electrode 12a can be predicted by equation (4).

[0376] After measuring the voltage on electrode 12a, system 100 can use inverse kinematics to calculate the charge density (and / or dipole density) on the atrial surface. Since the linear system is under-determined, a regularization algorithm can be used. In some embodiments, the scaling algorithm uses a weighted minimum norm estimate for regularization.

[0377] Weighted Minimum Norm Estimation (WMNE)

[0378] The scaling algorithm can use weighted minimum norm estimation (WMNE) to compensate for the distance effect by introducing a weighting matrix into the objective function:

[0379]

[0380] Where the weighting matrix W∈R n×n It is a diagonal matrix, and the diagonal elements serve as the norms of the columns of A.

[0381]

[0382] The solution to (5) is

[0383]

[0384] G = (W) T W) -1 =diag(1 / ||A(:,i)|| 2 ), i = 1, 2, ..., n, (6) can be rewritten as:

[0385]

[0386] The positive matrix can be rewritten as A = U∑V using singular value decomposition. T And by replacing (7), a simplified solution for WMNE is obtained.

[0387]

[0388] Back expansion

[0389] Once the charge density is calculated using the weighted minimum norm estimation regularization method, system 100 can calculate a post-scale value to adjust the charge density amplitude. First, system 100 uses a uniform charge distribution to calculate the voltage on the array 12 electrodes 12a. Then, a calibration charge density can be calculated using the voltage generated by the uniform source. The calibration charge density can be compared to the uniform charge density to calculate the post-scale value. Finally, the post-scale value can be used to adjust the charge density on the atrial surface.

[0390] By pre-weighting the measurement matrix to compensate for source depth and post-scaling the reconstructed source to eliminate systematic errors, the pre- and post-scaling techniques increase sensitivity to sources spatially distant from electrode 12a. These improvements have been demonstrated in both the applicant's simulations and clinical studies.

[0391] Improved high-pass filter with fewer artifacts

[0392] Biopotential signals (such as intracardiac electrograms and surface ECG signals) can contain unwanted signal components, such as DC offsets of varying amplitudes between channels, manifested as baseline spread, respiration, and low-frequency drift. These signal components typically consist of relatively low frequencies, e.g., frequencies <= 0.5 Hz, compared to the spectral content of cardiac activation morphology. High-pass filters (HPFs) are commonly used signal processing tools to remove or reduce low-frequency information from signals. However, conventional high-pass filters can also introduce artifacts into the morphology of EGM / ECG signals, which can affect the accuracy and resolution of 3D electrophysiological images. High-pass filters are often difficult to simultaneously provide high stopband attenuation, abrupt transition bands, and low signal distortion in a single design.

[0393] In some embodiments, the BIO filter 33, biosignal processor 34, or bioprocessor 36 may include an improved high-pass filter that reduces the aforementioned undesirable low-frequency signal components with high stopband attenuation, a sharp transition band, and low signal distortion. The low-pass filter is used to remove frequency components above a specified frequency, such as frequencies in the range of 0.01 to 2.0 Hz, for example, 0.5 Hz. The output of the low-pass filter is then subtracted from the original signal. The low-pass filter may include two stages with distinct cutoff frequencies and operate at different sampling rates. Compared to a standard high-pass filter, the two-stage arrangement allows for a larger transition band and lower distortion while maintaining low computational complexity, allowing the filter to be implemented in software or firmware. Compared to a standard high-pass filter, this filter preserves the baseline shape, signal amplitude, and flatness of the biopotential signal.

[0394] Based on inverse far-field source rejection

[0395] Far-field activations from sources outside the region of interest in a 3D electrophysiological image can be rejected by performing field calculations for artificially located far-field sources. As an example, measurements of activation on surface A of a cardiac chamber of interest (e.g., an atrium) may include contributions from sources in another chamber (e.g., a ventricle). Then, when mapping activation patterns in the atrium, activations from far-field sources in the ventricle may be inappropriately assigned to sources on surface A of the atrium. In some embodiments, system 100 may include a method (e.g., an algorithm) to help reject activations from far-field sources from an image of the chamber of interest by performing the following steps:

[0396] 1. Create an additional source region B (point, point set, or surface) and attach it to the anatomical surface A using anatomically accurate relative orientation, for example, the location of a valve when the chamber of interest is the atrium and the far-field source originates from the ventricle.

[0397] 2. Perform inverse solution on the union of surfaces A and B.

[0398] 3. Only display the image obtained from surface A.

[0399] i. Far-field sources that are essentially in common mode with the measured electrograph should be assigned to sources on surface B, thereby reducing any far-field artifacts on the image of the chamber of interest.

[0400] 4. Perform the forward solution from the source on surface A to produce only an electrical record without far-field components, for example, for inverse-based V-wave removal.

[0401] The method can reduce artifacts or residuals on images and electrographs caused by filters or template-based methods.

[0402] Surface reconstruction with connection structure

[0403] The reconstruction of anatomical chambers or surfaces described in International PCT Patent Application No. PCT / US2016 / 032017, filed May 12, 2016, entitled "Ultrasound Sequencing System and Method," may lack detail on connecting or adjacent structures. As an example, the reconstruction of the left atrium can accurately produce a representation of the atrial body, but attached venous structures may lack definition beyond the venous foramen. System 100 may include methods (algorithms) incorporating one or more of the following to improve the anatomical definition of connecting structures.

[0404] The method may include data processing of a surface point cloud to create a master surface (a surface reconstructed according to PCT / US2016 / 032017), which identifies regions including surface points from or near a connecting structure. In some embodiments, the data processor may include a spatial distribution of surface points within the region. In some embodiments, the data processor may include a spatial distribution of surface points within the region relative to the master surface.

[0405] The method may also include data processing of localization data sampled by devices within the chamber of interest, such as basket mapping catheters and / or auxiliary diagnostic catheters distributed throughout the entire volume of the chamber. In some embodiments, such localization data is provided from data recorded during a surface reconstruction process. In some embodiments, such localization data is provided from data recorded during independent acquisition. In some embodiments, field features can be approximated by one or more methods disclosed in the above description of intracardiac field estimation. In some embodiments, regions including and / or adjacent to surface points within / near and / or including the connecting structure are identified based on differences in the localization fields in these regions, differences inherently imposed by the anatomical presence of the connecting structure. In some embodiments, the principal direction vector describing the orientation of the connecting structure can be determined by: local differences in the estimated intracardiac localization field relative to the surrounding region, spatial distribution of surface points in or near the region, principal components of surface points in or near the region, surface features of the principal surface, position or orientation of surface points in the region relative to the principal surface, position or orientation of a set of field estimates from the sampled localization data, or a combination thereof.

[0406] The method may also include processing steps that reject consideration of surface points based on: local differences in the estimated intracardiac positioning field relative to the surrounding region, spatial distribution of surface points in or near the region, principal components of surface points in or near the region, surface features of the master surface, position or orientation of surface points in the region relative to the master surface, position or orientation of field estimation sets from the region, position or orientation of sampled positioning data sets of devices from the chamber of interest, range data from a range sensor set (e.g., an ultrasound transducer), or a combination of these.

[0407] The method may further include processing steps to reconstruct a representative surface or surface segment, comprising one or more identified regions from or near the connecting structure and / or including surface points of the connecting structure. In some embodiments, the representative surface is reconstructed from one or more identified regions of surface points, independent of information from other surface data. In some embodiments, the processing steps for the representative surface include information from a master surface. This process may be repeated for each connecting structure. In some embodiments, the reconstruction of the representative surface may utilize a meshing algorithm, such as a ball-pivot or alpha-shape algorithm. In some embodiments, the reconstruction of the representative surface may use an iterative fitting or deformation algorithm, such as an algorithm that performs the following: starting with a generalized surface shape, modifying one or more regions of the surface shape into a set of points, estimating a cost function such as a distance error, evaluating the cost function against a termination criterion such as a threshold, and iteratively continuing to determine the next modification to the surface shape if the termination criterion is not met. In some embodiments, the reconstruction of the representative surface may utilize an analytical set of spatial sequence basis functions, such as spherical harmonic expansion, to approximate the best-fit surface of the point set.

[0408] The method also includes processing steps to combine or merge surface segments of one or more representative surfaces and / or connection structures into a master surface or a portion thereof. In some embodiments, surface merging is performed simultaneously with the reconstruction of representative surfaces and / or surface segments of each connection structure. In some embodiments, a mesh blending procedure is used for surface merging. In one embodiment, the merging procedure may include one or more of the following steps:

[0409] 1. Use a mesh generation algorithm (e.g., alpha shape algorithm) to create a coarse mesh to unite two structures. The coarse mesh includes the main surface and the identified set of surface points from or near the connecting structure and / or including the connecting structure.

[0410] 2. Use a fitting algorithm (e.g., principal component analysis) to determine the best-fit shell for the identified surface points. The best-fit shell can have a simple shape, such as an ellipsoid, cylinder, sphere, or similar geometry.

[0411] 3. A subset of triangles in the coarse mesh is identified as openings or apertures in the connecting structure at the intersections with the best-fit shell.

[0412] 4. The portion of the coarse mesh that extends beyond the orifice is retained and connected to the main surface via triangles along the orifice.

[0413] 5. Using the surface normal of the triangle at the orifice as a guide, remove the inner triangle based on the continuous surface normal direction.

[0414] 6. Seal the holes in the resulting surface.

[0415] 7. Smooth the resulting surface, remesh it, or do both.

[0416] Now for reference Figure 19 This invention illustrates a method for recording cardiac activity and displaying an anatomical model of a "beating" heart, consistent with the inventive concept. Method 1900 includes: recording cardiac activity data over multiple cardiac cycles, including electrical activity data and wall position data; analyzing the electrical activity data into discrete segments of the determined recorded cardiac cycles; segmenting the wall position data by discrete segments; generating an anatomical model of the heart chambers for each discrete segment component; and displaying the beating anatomical model based on the recorded cardiac activity data.

[0417] In step 1910, cardiac activity data is recorded, such as data recorded using one or more recording devices of system 100 (e.g., catheter 10), as described herein. Electrode array 12 may include one or more electrodes 12a and / or one or more ultrasound transducers 12b. System 100 may be configured to position electrode array 12 within the cardiac chamber (i.e., determine the position and orientation of array 12 relative to a positioning field generated within the patient's body) and record the patient's cardiac activity data. Cardiac wall position data ("anatomical" data) may be recorded via one or more ultrasound transducers 12b, as described herein. Simultaneously (or relatively simultaneously, e.g., in an interleaved mode), cardiac electrical activity data ("biopotential" data) may be recorded via one or more electrodes 12a, also as described herein. In some embodiments, data is recorded for a period of time (e.g., over multiple cardiac cycles), and after a complete set of data has been recorded, the data is manipulated as described below, and a beating anatomical model is generated and displayed as a "recorded" video. Alternatively or additionally, a first subset of data is recorded and manipulated (e.g., analyzed), a subsequent second subset of data is recorded, and a pulsating anatomical model can be generated and displayed based on the first and second subsets of data. In this configuration, a "real-time view" of the pulsating anatomical model (e.g., a view with a short delay between actual anatomical motion and the corresponding video representation) can be displayed.

[0418] In step 1920, for example by Figure 1 The processor of console 20 analyzes the recorded biopotential data to determine time periods or segments representing various phases of the cardiac cycle (e.g., heartbeat phases). For example, recorded biopotential data of regular and / or near-regular heart rhythms (e.g., sinus rhythm or fibrillation) can be segmented into periodic groups of time periods representing phases of the cardiac cycle. In some embodiments, these segments repeat with small variations between cycles. In some embodiments, the biopotential data is analyzed to determine key time indices (e.g., the rise or fall of the V wave) for unique phases of the cardiac cycle, and the data between each identified consecutive key time indices is segmented into a predetermined number of variable duration segments. Alternatively, the data between each consecutive key time indices is segmented into a variable number of predetermined duration segments.

[0419] In step 1930, the recorded anatomical data is aggregated according to the time periods determined in step 1920. For example, for each common time period of each recording cycle, the recorded anatomical data (including a set of points representing at least a portion of the cardiac surface during each time period) is aggregated to form an anatomical data set, which includes points from multiple cycles representing a common time period (e.g., a common phase of the cardiac cycle). This process is performed for each time period for each recording cycle and compiled into a data set, such as a data set including multiple "boxes". The total number of boxes can be set to be equal to the average number of time periods of the recorded cycle. In some embodiments, such as when the biopotential data is segmented into defined duration segments, consecutive cycles can include more or fewer segments than the number of boxes, and the data can be adjusted to eliminate differences. For example, if a cycle includes more than the number of boxes, additional segments can be skipped (e.g., ignored). Additional segments can be skipped by removing them from the "end" of the cycle, and / or an algorithm can be used to determine which segments to skip, such as skipping segments 11, 22, and 33 if three additional segments exist in a cycle including 33 segments. Additionally or alternatively, additional time periods may be "assigned" to different bins, such as the last bin or the last two bins of a cycle. In some embodiments, the bins to which the additional time periods are assigned are jittery.

[0420] In some embodiments, if the period comprises fewer than the number of bins, the time periods can be distributed across a subset of bins. The distribution can be sampled uniformly (e.g., earlier segments are “assigned” to earlier bins), or segments can be analyzed and assigned to bins based on having similar properties to the data in the bins to which they are to be assigned (e.g., data that already exists).

[0421] In step 1940, for each bin, the data contained therein is used to create an anatomical model of the cardiac chambers for the relevant time period of the represented cardiac cycle. In some embodiments, the anatomical model is created as described herein, and / or with reference to co-pending U.S. Patent Application No. 15 / 128,563, filed September 23, 2016, entitled "System and Method for Cardiac Analysis User Interface," which is incorporated herein by reference in its entirety for all purposes. In some embodiments, a "base model" is created, which includes a model created from a single bin representing a single time period, such as a time period representing cardiac systole or diastole. Anatomical models representing the remaining bins can be generated based on the base model, for example, by using a conformal mesh. Each vertex of the base model can be projected inward and outward along its normal axis, and each vertex can be adjusted along the projection to match and / or at least approximate the data in subsequent bins, for example, to create an anatomical model representing the data in each bin based on the base model. One or more smoothing filters can be used to improve (e.g., smooth) the created anatomical model. For example, based on the distance from a fixed location (e.g., the origin) to each vertex of each anatomical model, a low-pass filter can be used to adjust each vertex in each anatomical model. When the anatomical model is displayed as a dynamic model (also referred to herein as a beating heart model), such filtering can produce "smoother" and / or more coherent motion, for example, in step 1950.

[0422] In step 1950, a continuous anatomical model representing each phase of the cardiac cycle is displayed as a dynamic model. In some embodiments, electrical data, such as calculated data based on recorded cardiac activity data as described herein (e.g., voltage data, dipole density data, and / or surface charge data), are superimposed on the dynamic model. The superimposed image allows the operator to visualize one or more features of the recorded anatomical structures, such as electromechanical coupling and / or delay. In some embodiments, system 100 processes the data representing the dynamic model to determine one or more measures relating to electrical activity and / or the size, shape, and / or motion of anatomical structures. For example, system 100 may measure and / or display one, two, or more of the following: ejection fraction, stroke volume, chamber volume, chamber size, regional wall displacement, velocity, or other indices / representations of stiffness or loss of function. System 100 may calculate and display certain physiological parameters over time, such as chamber volume (e.g., atrial volume). For example, atrial or other chamber volumes may be presented (e.g., as curves) in a graph representing chamber volume versus time. Functional characteristics, such as displacement or velocity of the walls at each vertex or region, can be displayed (instantaneously within a time window) as an image (e.g., a color graph) depicting a representation of contractility (e.g., an index). Differences and / or changes in functional characteristics before and after a clinical event (e.g., ablation surgery) can similarly be displayed as a color graph depicting the altered function of the region caused by the clinical event. Additionally or alternatively, one or more surface dynamics, such as tissue stress or strain, blood velocity, regional wall contractility, or other indices / representations of stiffness or loss of function, can be measured and / or displayed.

[0423] System 100 can be configured to determine metrics related to the size, shape, and / or motion of a patient's heart, for example, by creating a dynamic model as described above. The evaluation of these metrics can be performed by System 100 as part of a safety algorithm (e.g., processor 26 or other components of console 20). The safety algorithm can run continuously, or at least during procedures performed on the patient, such as cardiac surgery using a cardiac ablation device (e.g., an endocardial ablation catheter or epicardial ablation device). The cardiac ablation device can be connected to console 20 (e.g., when console 20 is configured to supply energy to the ablation device) or an energy delivery device operatively connected to console 20 (e.g., enabling console 20 to control the energy delivery to the cardiac ablation device). When an unsafe or undesirable condition is detected, System 100 can issue warnings (e.g., via auditory, visual, and / or tactile user output components of user interface 27) and / or adjust (e.g., stop or at least reduce) the operation of the cardiac ablation device. For example, system 100 may be configured to detect undesirable conditions selected from the group consisting of: tamponade (e.g., due to pericardial hemorrhage); stroke; deviation from stable hemodynamic conditions; reduced ejection fraction; reduced stroke volume; reduced wall motion; undesirable decrease in chamber volume (e.g., undesirable decrease in mean chamber volume); undesirable increase in chamber volume (e.g., undesirable increase in mean chamber volume); and combinations thereof. Undesirable conditions can be conditions of global cardiac chamber characteristics (i.e., the entire chamber) or regional ventricular characteristics (e.g., a portion of the ventricular wall). In some embodiments, the undesirable condition is a reduction in atrial volume (e.g., a reduction in atrial volume at a comparable heart rate), indicating reduced atrial filling, which may be associated with hemorrhage leading to tamponade. In some embodiments, system 100 is configured to detect tamponade (e.g., as described above) and alert the user and / or deliver (or prompt delivery) a coagulant or other treatment (other medication or interventional therapy) to control hemorrhage and / or tamponade.

[0424] In some embodiments, one or more continuous anatomical models may be improved by the user, for example using mesh generation and / or smoothing techniques, as referenced below. Figure 21 As described.

[0425] Now for reference Figure 20 This invention illustrates a method for recording cardiac activity data and updating an anatomical model of the heart (e.g., the chambers of the heart) based on "snapshots" of the data, consistent with the inventive concept. Method 2000 includes: recording cardiac activity data and generating an anatomical model of the heart; recording "snapshots" of the cardiac activity data, including wall position data; filtering the recorded data; determining approximate sizes and shapes of the recorded anatomical structures based on the snapshots of the data; and updating the anatomical model based on the approximate size and shape of the anatomical structures.

[0426] In step 2010, cardiac activity data is recorded, such as data recorded using one or more recording devices of system 100 (e.g., catheter 10), as described herein, and an anatomical model of the heart is generated. Electrode array 12 may include one or more electrodes 12a and / or one or more ultrasound transducers 12b. System 100 is configured to position the electrode array 12 within the cardiac chambers (i.e., determine the position and orientation of array 12 relative to the positioning field generated within the patient's body) and record the patient's cardiac activity data. Cardiac wall position data ("anatomical" data) may be recorded via one or more ultrasound transducers 12b, as described herein. Simultaneously (or relatively simultaneously, e.g., in an interleaved mode), cardiac electrical activity data ("biopotential" data) may be recorded via one or more electrodes 12a, also as described herein. A first set of data may be recorded over a first time period, such as at least 5 seconds, 15 seconds, or 30 seconds, or less than 1 minute or less than 5 minutes. A static anatomical model of the heart (e.g., the cardiac chambers from which data is recorded) is generated using a first set of recorded data, such as a model representing the chamber shapes during cardiac contraction, diastole, and / or intermediate shapes between these two phases. The method for generating the static model is described herein and also in co-pending U.S. Patent Application No. 15 / 128,563, filed September 23, 2016, entitled "Heart Analysis User Interface System and Method," which is incorporated herein by reference in its entirety for all purposes. In some embodiments, the anatomical model is generated from CT and / or MRI data recorded prior to and / or during method 2000. Alternatively, the anatomical model may include a generic anatomical model configured to be updated by method 2000. In some embodiments, the static anatomical model may be improved by the user, for example using mesh generation and / or smoothing techniques, such as those referenced below. Figure 21 As described.

[0427] In step 2020, a second set of data may be recorded during a second time period (e.g., a time period shorter than the first time period). In some embodiments, the second set of recorded data represents an instantaneous "snapshot" of the cardiac chamber, for example, when the second time period includes a time period of less than 0.5 seconds, such as less than 100 ms, or less than 30 ms. The second set of recorded data may include sparse sampling of the cardiac chamber, for example, including sampling of fewer than 96 samples (i.e., fewer than 96 ultrasound measurements). The second set of recorded data may include multiple ultrasound measurements performed simultaneously and / or nearly simultaneously in multiple directions by the electrode array 12, for example, to allow approximation of the size and / or shape of the anatomical structure from sparse sampling of the anatomical structure, as described below.

[0428] In step 2030, the recorded second set of data may be filtered and / or manipulated. In some embodiments, one or more ultrasound reflections may be missing from the data set (e.g., if a reflection is not detected in response to an ultrasound pulse or "ping"), and system 100 may be configured to "fill" the missing data, as described below. In some embodiments, fixed values ​​(e.g., fixed values ​​determined by system 100) may be inserted to minimize the negative impact of subsequent calculations on the data set, such as fixed values ​​representing "middle" or neutral (neither systolic nor diastolic) locations of anatomical structures. In some embodiments, fixed values ​​based on historical averages (e.g., based on the average from the first set of data) may be inserted to replace missing data points. In some embodiments, fixed values ​​based on a single historical value may be inserted, such as the "most recent" valid data point (e.g., a valid data point from the most recently recorded transducer with missing data). In some embodiments, missing data may be inserted using a fitting algorithm, such as an interpolation algorithm, such as an algorithm using linear, spline, piecewise continuous, and / or other interpolation methods. In some embodiments, missing data may be inserted based on values ​​determined by adaptive and / or machine learning algorithms, such as algorithms based on one or more subsets of valid data within the data set.

[0429] In step 2040, the size and / or shape of the anatomical structure represented by the second set of recorded data is determined. In some embodiments, an "envelope" algorithm (e.g., convex hull) may be implemented to determine a rough estimate of the size and shape of the anatomical structure by encapsulating all measured ultrasound points within a complex shape. Alternatively or additionally, a stress / strain model may be applied to the static model generated from the first set of data, and a push / pull algorithm may be applied to the stress / strain model using the second set of data, for example, to find a "fit" between the stress / strain model and the second set of data at "instantaneous" points.

[0430] In some embodiments, the static model can be composed of a vertex mesh P 原始 (x, y, z) represents that it can be "updated" to P by the following formula. 瞬间 (x, y, z):

[0431] P 原始 (x, y, z) = P 瞬间 (x, y, z) + ΔP

[0432] The second set of recorded data was used to estimate ΔP, and P 原始 The variation of each vertex of (x, y, z) is limited to the direction of the local normal. ΔP can be limited to a combination of functions, i.e., functions with weighted parameters λ. n Global anatomical structure fitting, regional anatomical structure fitting, temporal and spatial smoothing:

[0433] ΔP=λ1(F 全局拟合 )+λ2(F 区域拟合 )+λ3(F 空间平滑 )+λ4(F 时间和空间平滑 )

[0434] For example, a global fitting function can be achieved by separating each ultrasound data point from the mean or median of its nearest location on the anatomical structure. Region fitting can be used to establish a relationship between each ultrasound point and a vertex on the anatomical structure. This relationship can be established by defining a function (e.g., inverse distance or Gaussian) between all ultrasound points in a second set of data and the vertices of a static model. Once this relationship between measurements and the anatomical structure is established, the anatomical structure can be modified. The λ values ​​weight different aspects of the solution, or they can themselves be functions of the vertices and can be set based on the confidence level of the ultrasound measurements or the distance to the nearest ultrasound measurement.

[0435] In some embodiments, the static model can be composed of vertex grid P 原始 This means that it can be "updated" to P by the following formula. 瞬间 (t):

[0436] P 瞬间 (t)=P 原始 +ΔP(t), (1)

[0437] ΔP is estimated using all or a subset of instantaneous ultrasound data points (the second set of data points). The formula in Equation (1) allows a static grid to become a variable grid, where the deformation is due to variations in ultrasound range data related to changes in the location of the heart wall (and therefore contractility and volume). In some embodiments, there will be fewer ultrasound data points than at the apex, thus requiring estimation of ΔP(t) at most locations.

[0438] The unknown ΔP(t) can be estimated using interpolation and / or extrapolation methods, energy function minimization, and / or data models such as stress-strain relationships. Interpolation and / or extrapolation methods can include inverse distance weighting, nearest neighbor inverse distance weighting, basis function interpolation, and / or kriging. Basis function interpolation implementations can include defined basis functions (e.g., spherical harmonic basis functions) and use these functions to estimate unknown data points. The energy (cost function) formula can also be used to determine ΔP(t) that minimizes the instantaneous energy.

[0439]

[0440] The weight w can be used. n To limit different energy levels E n This represents the relative influence of the energy term on the total energy. The weighted term w n It can be a function of other measurable quantities. An example of an energy term could be E. 数据拟合 This can be defined as the error between the ultrasound range data and the corresponding vertices on the anatomical structure. Another energy term can be defined as E. 原始形状 This can be defined as the deviation from the spatial gradient of the original anatomical structure. The total energy having these terms can then be defined as...

[0441]

[0442] Model-based deformation methods can include stress / strain models that define the anatomical structure. Given the deformation, stress / strain can be estimated at each location. Stress / strain can then be interpolated, or an energy function can be established and minimized to determine the stress and strain at each vertex. Vertex positions can then be updated to correspond to the estimated stress / strain. Once ΔP(t) has been calculated for each vertex and time point, several metrics can be computed, such as instantaneous blood volume, active and passive blood volume emptying, and / or region-based contractility.

[0443] In some embodiments, the interpolated deformable mesh is configured as follows: For each time moment, the coordinates of the ultrasound transducers within the cardiac chamber are determined. For each time moment, a vector vec is determined for each ultrasound transducer. n Its limited ultrasound range data USpos n And the M-mode position line of the ultrasound transducer. Determine the intersection point between the M-mode ultrasound transducer and the static cardiac anatomy. Determine the closest ultrasound vector (vec). n The vertex of the intersection of the anatomical structure and the vertex n Calculate the ultrasound range data USpos n and vertex n Distance between n Assign ΔP(n,t) to dist n The projection and at the vertex n The local surface normals are determined. Once all ΔP(n,t) of the valid ultrasound data are determined, interpolation for the remaining vertices is determined using inverse distance scaling with nearest neighbors. For each time step, the volume of the anatomical structure is calculated using the updated vertex positions.

[0444] In step 2050, based on the first data set and updated by the second data set, a dynamic anatomical model can be displayed, similar to the reference above. Figure 19 The steps described in 1950.

[0445] Now for reference Figure 21 This illustrates a method for generating and editing an anatomical model of the heart, consistent with the inventive concept. Method 2100 includes: recording a first set of cardiac activity data, including wall position data; creating a basic anatomical model of the heart (e.g., the chambers of the heart) based on the first set of cardiac activity data; collecting additional cardiac wall position data and updating the basic anatomical model; and editing the basic anatomical model, for example, via a graphical user interface.

[0446] In step 2110, a first set of cardiac activity data is recorded, such as data recorded using one or more recording devices of system 100 (e.g., catheter 10), as described herein. Electrode array 12 may include one or more electrodes 12a and / or one or more ultrasound transducers 12b. System 100 is configured to position electrode array 12 within the cardiac chamber (i.e., determine the position and orientation of array 12 relative to a positioning field generated within the patient's body) and record the patient's cardiac activity data. Cardiac wall position data ("anatomical" data) may be recorded via one or more ultrasound transducers 12b, as described herein. Simultaneously (or relatively simultaneously, e.g., in an interleaved mode), cardiac electrical activity data ("biopotential" data) may be recorded via one or more electrodes 12a, also as described herein. As used herein, anatomical data may refer to any data collected by system 100 (e.g., via electrode array 12 or other data collection devices of system 100) for generating an anatomical model. Anatomical data may include data from the following groups: data on surface points (e.g., surface locations) measured by ultrasound; volume data determined using positioning methods, such as tracking the location of electrodes within a chamber (e.g., a cardiac chamber that includes at least the volume of a bulge containing the collected electrode locations) over time; data on surface points collected using positioning methods, such as data collected via a contact sensing catheter of system 100 used with a positioning system; data acquired from outside the body (e.g., using external visualization devices), such as MRI and / or CT data; and combinations of one or more of these.

[0447] In step 2120, a basic anatomical model of the heart (e.g., including a set of vertices and polygons defining a 3D surface) is generated based on the first set of recorded data. As described below, method 2100 is used to generate and edit static anatomical models of the heart (e.g., cardiac chambers from which data is recorded), such as models representing the chamber shapes during cardiac contraction, cardiac diastole, and / or intermediate shapes between these two phases. The method for generating the static model is described herein and also in co-pending U.S. Patent Application No. 15 / 128,563, filed September 23, 2016, entitled "Heart Analysis User Interface System and Method," which is incorporated herein by reference in its entirety for all purposes. Additionally, method 2100 can be iterated (e.g., by iterating through a series of static anatomical models representing successive phases of the cardiac cycle) to generate and edit dynamic models of the chambers, such as the dynamic models described herein. In some embodiments, the basic anatomical model includes an anatomical model imported into system 100 from an external imaging source (e.g., MRI and / or CT).

[0448] In some embodiments, the static anatomical model comprises a model generated using methods that "build" complex 3D geometries around the collected point sets during or after collection, such as by using alpha-shape methods and / or "growth" methods. For example, a static model of the chambers can be constructed during cardiac diastole (i.e., the heartbeat phase when the myocardium relaxes and allows the chambers to fill with blood, and the chambers are at their maximum volume) using the growth method described below. In these embodiments, when the first set of data is recorded, the point set is considered to exist only within the chambers, and throughout all phases of the heartbeat, the fitting algorithm (e.g., convex hull, including all (excluding outliers) collected points) can be considered to represent the chambers at their maximum volume (i.e., cardiac diastole). In some embodiments, a similar approach can be used, and a gating mechanism (e.g., a gating mechanism based on cardiac electrical activity data) can be included to include only points collected during specific phases of the heartbeat (e.g., during cardiac systole, when the chambers are at their minimum volume).

[0449] In step 2130, a second set of cardiac activity data ("additional" data) is recorded, and the base anatomical model can be updated based on the additional data. In some embodiments, such as when the base anatomical model includes MRI and / or CT images, the additional data can be used to register the imported model to anatomical structures within the coordinate system of system 100. Additionally or alternatively, the additional data can be used to "fill in" or increase the resolution of regions of the anatomical model. For example, system 100 can (e.g., via a user interface) inform the user of regions with low data density (e.g., sparsely sampled regions). The user can collect additional data associated with such regions, for example, by "pointing" (e.g., turning) the catheter 10 to the indicated region, and record the additional data. The collected additional data can be compiled using the first set of data to generate an improved base anatomical model. Alternatively, the additional information can be used to update the base anatomical model. In some embodiments, the user can resample regions, for example, when the user suspects a problem with the model. The additional data can be used to verify the accuracy of the base anatomical model and / or update the base anatomical model.

[0450] In step 2140, system 100 may provide one or more manual and / or automated tools for editing the basic anatomical model, for example, via a graphical user interface and one or more user input devices, such as a mouse, keyboard, touchscreen display, etc. As described below, user "actions" are performed using controls, programs, and / or algorithms included in and / or enabled by system 100. In some embodiments, it may be advantageous to add one or more structures (e.g., "anatomical projections") to the basic anatomical model, for example, to represent the left atrial appendage and / or pulmonary veins. System 100 may allow users to add "simple" geometric projections (e.g., projections with a basic shape, along a straight or approximately straight path, or projections with little or no change in the basic shape along the projection), and / or more complex geometric projections, such as projections formed by adding "stress" along a user-defined vector to deform the basic anatomical model, both of which are defined below.

[0451] In some embodiments, the user defines the projection to be added to the base anatomical model by inputting the intersection points of the projection, the vector (direction and length) of the projection, and the shape of the projection. Intersection points can be selected by selecting vertices of the base anatomical model. The user can define the projection vector by orienting the base anatomical structure at a specific location on the screen (e.g., by rotating the model so that the desired vector is in an "up" direction, or "into" the screen, etc.), by "drawing" the desired vector (e.g., by creating a line in 3D space by selecting a first and a second point), or by any other method of defining the vector. The user can adjust the length and direction of the vector, for example by "dragging" the endpoint of the vector along its direction (to change the length of the vector), or by freely "dragging" the endpoint of the vector in space (to change the direction and / or length of the vector). The user can then define the basic shape to be projected along the vector by drawing a shape that highlights a portion of the base anatomical model, or by selecting a portion of the base anatomical model (e.g., one or more vertices and / or polygons). The user can use one or more drawing tools provided by System 100 to define the basic shape, such as drawing a circle, ellipse, or other defined shape, or a free-form shape. The drawn shape can be projected along a vector to select the vertices of the anatomical model enclosing the projection, and these vertices can be projected along a vector to form the final projection to be added to the base model. In some embodiments, the projection follows a user-defined shape (e.g., the system does not limit the projection to the enclosing vertices), and the system performs a fitting function (e.g., fillet) to approximate the transition between the base model and the projection. In some embodiments, the user can define the base of the projection by using a coarse selection tool to select (or deselect) the vertices of the base model using a "brush" function. Additionally or alternatively, the coarse selection tool may include a "lasso" type editing function that automatically selects the vertices enclosing the selected vertex group. Additionally or alternatively, the user can define the base of the projection by individually selecting the vertices and / or polygons of the base anatomical model, for example, using a fine selection tool.

[0452] In some embodiments, projecting the vertices of the underlying anatomical model allows the system to process (e.g., refine) the improved anatomical model into a single "mesh" (a group of polygons defining a 3D surface), such that the projection is processed together with the remaining underlying model of the anatomical structure. Processing may include, but is not limited to: smoothing, re-dividing, subdivision, hole closure, isolation, or floating triangle identification and removal. Additionally or alternatively, the projection may be processed separately from the remaining underlying model, such that the user-defined structure is not modified when processing the improved anatomical model.

[0453] In some embodiments, system 100 allows a user to perform one of the following actions: confirm a created projection (e.g., save the projection to an improved model); clear the created projection and / or the anatomical points used to create the projection, and create a new projection; define the "behavior" of the projection relative to electrical activity on the model (e.g., treat the projection as a "hole" in the anatomical model without displaying electrical activity); modify the geometric parameters of the projection, such as length, scaling of cross sections, intersections, and / or vector direction of structures; delete previously created structures; erase anatomical points; remove portions of the projection and / or other portions of the anatomical model; modify regions of vertices or polygons by "push or pull" elements using a user input device (e.g., a mouse); and combinations of one or more of these.

[0454] One or more steps for creating an improved anatomical model can be performed semi-automatically and / or automatically (“automatically” herein). In some embodiments, automatic modifications to the base model by System 100 can be overturned by the user. In some embodiments, one or more parameters defining the projection (e.g., its orientation, shape, size, length, cross-section, position, and / or intersection) can be defined by System 100 based on one or more inputs (e.g., a first set of anatomical data). In some embodiments, as described above, the user can collect additional data points to update the anatomical model, and System 100 can update the anatomical model automatically, for example, by adding a projection to the model. Based on incremental data acquisition (e.g., based on the rate at which System 100 collects data and determines additional data to ensure model updates), System 100 can update the model based on the additional data at a rate from once every 1 millisecond to once every 10 seconds, for example, once every 100 ms or every 500 ms. System 100 can determine whether the additional data is related to the projection based on one or more characteristics of the data. For example, the additional data itself may include a set of data that is primarily associated with a segment of an anatomical structure, which is best modeled as a projection from a base anatomical model, and the data represents a shape with one or more principal components (e.g., a shape that approximates the anatomical structure).

[0455] System 100 can be configured to provide one or more principal components and determine one or more properties of the projection to be added to the anatomical model. For example, the principal component vectors of the data can be used to determine the principal axes of the structure. The intersection of the projections can be determined by the intersection of the principal direction vectors with the base anatomical model. The shape of the projection can be determined by the cross section, outer boundary, or other measure of the supplementary data. Methods for determining the cross section or outer boundary of the projection parameters can utilize a plane perpendicular to the base model at the intersection and / or perpendicular to the direction vectors of the supplementary data to determine the shape of the projection. In some embodiments, these methods can utilize a subset of vertices of the base model within a threshold distance from the intersection to define the projection. In some embodiments, these methods can utilize only a subset of the supplementary data determined to be "outside" the base model. In some embodiments, the projection can be based on the dot product of the vector between a point in the supplementary data and the nearest vertex of the base model and the normal vector from the vertex of the base model.

[0456] In some embodiments, system 100 adds projection using the boundary method described below. A boundary surface is determined using a convex hull that defines the set of all anatomical data "points" (e.g., from a first set of data and any additional data collected). The boundary surface is determined using an α-shape that defines the set of anatomical points. Internal vertices and triangles are removed from the data, and subsets of data may be excluded from the projection to be created, such as subsets of points inside the underlying anatomical structure or subsets of points within a distance of the surface of the underlying anatomical structure. A subset of vertices ("patches") of the underlying anatomical structure that intersect the boundary surface is removed from the underlying anatomical structure, creating holes and / or boundaries connected to the created structure. The boundaries of the holes created on the underlying anatomical structure and the edges of the created structure may be blended to produce a "watertight" boundary (e.g., a seamless boundary). In some embodiments, system 100 may compute a new α-shape from the improved anatomical structure to create a new improved anatomical structure that fills any potential holes created by adding projection. In some embodiments, the automatically generated anatomical structure may be dynamically changed or updated by (e.g., by a user) acquiring additional data and / or erasing vertices or data points.

[0457] While the preferred mode and / or other preferred embodiments have been described above, it should be understood that various modifications may be made therein, and the invention can be implemented in various forms and embodiments, and they can be applied in many applications; only some of them have been described herein. The appended claims are intended to claim protection for the literal description and all its equivalents, including all modifications and variations falling within the scope of each claim.< / double> < / double> < / double>

Claims

1. A dynamic display system for cardiac information, comprising: One or more sensors configured to record cardiac activity data over multiple cardiac cycles, including electrical activity data or bioelectrical potential data, and wall position data; and The cardiac information console includes: The signal processor is configured as follows: Processing electrical activity data or biopotential data to determine discrete phases representing multiple cardiac cycles, wherein the signal processor analyzes the biopotential data to determine the time period or segment representing each phase of the cardiac cycle. Segmented wall position data are grouped into data sets, each data set comprising a set of points representing at least a portion of the cardiac surface during one of the discrete phases; and Generate anatomical models for each discrete stage in the data set; and A user interface module configured to display a beating anatomical or dynamic model of at least one cardiac chamber, wherein a continuous anatomical model representing each phase of the cardiac cycle is displayed as a dynamic model. The signal processor is configured to process electrical activity data to determine time indices associated with a unique phase of the cardiac cycle, and to segment the data between each consecutive time index into a set number of time periods.

2. The system according to claim 1, wherein, One or more sensors include at least one biopotential electrode and at least one ultrasonic transducer.

3. The system according to claim 2, wherein, One or more sensors are multiple biopotential electrodes and / or at least one ultrasonic transducer is multiple ultrasonic transducers.

4. The system of claim 1, further comprising an expandable electrode array, the expandable electrode array comprising a plurality of biopotential electrodes and a plurality of ultrasonic transducers.

5. The system according to claim 1, wherein, One or more sensors are configured to record cardiac electrical activity data and wall position data in an interleaved mode.

6. The system according to claim 1, wherein, One or more sensors configured to record cardiac activity data record bioelectric potential data, which is segmented into periodic time groups representing phases of the cardiac cycle.

7. The system according to claim 1, wherein, The key time index for the only phase of the cardiac cycle is the rise or fall of the V wave.

8. The system according to claim 1, wherein, The signal processor is configured to aggregate wall position data and segment the aggregated wall position data into multiple bins, each bin containing anatomical data from one of the time periods.

9. The system according to claim 8, wherein, The signal processor is configured to set the total number of bins to be equal to the average number of time periods of the multiple recorded cardiac cycles.

10. The system according to claim 8, wherein, When the time interval is more or less than the number of boxes, the signal processor is configured to adjust the recorded cardiac activity data to eliminate discrepancies.

11. The system according to claim 10, wherein, If the number of time periods exceeds the number of boxes, the signal processor is configured to skip the additional time periods.

12. The system according to claim 11, wherein, The signal processor is configured to remove the additional time period from the end of the cardiac cycle.

13. The system according to claim 11, wherein, The signal processor is configured to apply an algorithm to determine the time period to be skipped and the algorithm applies a pattern based on the number of additional time periods.

14. The system according to claim 11, wherein, The signal processor is configured to assign one or more additional time periods to different boxes.

15. The system according to claim 10, wherein, If the number of time slots is less than the number of boxes, the signal processor is configured to distribute the time slots into subsets of the boxes.

16. The system according to claim 14, wherein, The signal processor is configured to sample time periods uniformly.

17. The system according to claim 14, wherein, The signal processor is configured to assign time periods to bins that have characteristics similar to the cardiac activity data in the bins to be assigned.

18. The system according to claim 8, wherein, The signal processor is configured, for each chamber, to use wall position data contained in the chamber to create an anatomical model of at least one cardiac chamber representing a relevant time period of the cardiac cycle.

19. The system according to claim 8, wherein, The signal processor is configured to create an anatomical model of cardiac activity from a single bin representing a single time period.

20. The system according to claim 19, wherein, The signal processor is configured to generate an anatomical model representing the remaining bins based on a base model using a conformal mesh.

21. The system according to claim 20, wherein, The signal processor is configured to project each vertex of the base model inward and outward along its normal axis, and adjust each vertex along the projection to match and / or at least approximate cardiac activity data in subsequent bins, in order to create an anatomical model representing cardiac activity data in each bin based on the base model.

22. The system according to claim 21, wherein, The signal processor is configured to use one or more smoothing filters to improve the created anatomical model.

23. The system according to claim 22, wherein, The signal processor is configured to use a low-pass filter to adjust the position of each vertex in each anatomical model based on the distance from a fixed location and / or the origin of the coordinate system to each vertex of each created anatomical model.

24. The system according to claim 1, wherein, The cardiac information console is configured to superimpose electrical activity data onto the beating anatomy model.

25. The system according to claim 24, wherein, Electrical activity data includes surface charge data and / or dipole density data.

26. The system according to claim 24, wherein, The cardiac information console is configured to determine one or more measures relating to electrical activity data, including the size, shape, and / or motion of one or more of at least one cardiac chamber, superimposed on a beating anatomical model.

27. The system according to claim 24, wherein, The cardiac information console is configured to measure and / or display one or more of the following: ejection fraction, stroke volume, chamber volume, chamber size, region wall displacement, and / or velocity.

28. The system according to claim 24, wherein, The cardiac information console is configured to calculate and display atrial and / or other chamber volumes over time.

29. The system according to claim 24, wherein, The cardiac information console is configured to display a color graph with features depicting contractility representations, including one or more displacements and / or velocities of the cardiac wall at each vertex.

30. The system according to claim 29, wherein, The cardiac information console is configured to display differences and / or changes in the displacement and / or velocity of the heart wall before and after an event as a color map depicting changes in the region of altered functional characteristics caused by said event.

31. The system according to claim 24, wherein, The cardiac information console is configured to measure and / or display one or more surface dynamics, including one or more of tissue stress or strain, blood velocity, and / or regional wall contractility.

32. The system according to claim 1, wherein, The cardiac information console is configured to detect unwanted conditions selected from the following groups: tamponade; stroke; deviation from stable hemodynamics; reduced ejection fraction; reduced stroke volume; reduced wall motion; unwanted decrease in chamber volume; unwanted increase in chamber volume; and combinations thereof.

Images