Implantation guidance system

Abstract

A system may display a graphic of a heart that provides an indication of a target zone and an avoidance zone for implantation of the lead. The system may identify a location of the lead based on the location electrocardiogram, display on the graphic an indication of identified location; and determine whether the identified location is at an appropriate implantation site based on the location electrocardiogram. The system may, when the identified location is an appropriate implantation site, output an indication that the lead is at an appropriate implantation site. The system may, for each of one or more of tunneling electrocardiograms, determine whether the lead is at an effective position within the heart wall based on the tunneling electrocardiogram; and when the lead is at an effective position, output an indication that the lead is at an effective position.

Description

IMPLANTATION GUIDANCE SYSTEMCROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims the benefit of priority to U.S. Application No. 63 / 656,478 filed on June 5, 2024, which is hereby incorporated by reference in its entirety.BACKGROUND

[0002] Cardiac physiologic pacing (CPP) refers to any type of cardiac pacing intended to restore or preserve cardiac ventricular synchrony. This includes conduction system pacing (CSP), which is defined by pacing the intrinsic conduction system including either His bundle pacing (HBP) or left bundle branch area pacing (LBBAP). Cardiac physiologic pacing also includes cardiac resynchronization therapy (CRT), which involves the use of a left ventricular lead to pace the left ventricle and restore synchrony.

[0003] Cardiac dyssynchrony may be caused by injury or degeneration of the cardiac conduction system including the His bundle, left bundle branch (LBB), and right bundle branch. This dyssynchrony can lead to the abnormal pumping of blood that may be characterized by a low ejection fraction. This abnormality can have various adverse effects such as worsening heart failure symptoms, reduced exercise capacity, and a poorer quality of life. CPP involves the implantation of a pacemaker that sends electrical impulses to the left ventricle or right ventricle or both prompting them to contract synchronously. This improved coordination enhances the heart’s ability to pump blood effectively which can help to mitigate the adverse effects.

[0004] A pacemaker is a device that is designed to manage and maintain the heart’s rhythm by delivering electrical pulses to the heart wall. During CPP, a pacemaker is employed to treat patients who have an arrhythmia, for example, resulting from a complete heart block (i.e., a third-degree atrioventricular (AV) nodal block or left bundle branch block.) A complete heart block is a serious condition in which electrical impulses that control the heart’s rhythm are blocked from traveling from the atria to the ventricles.

[0005] A pacemaker has a pulse generator and may have multiple leads which may additionally have multiple electrodes that can be positioned at various locations on the heart wall. A lead may be unipolar with one electrode, bipolar with two electrodes, or multiple polar having three or more electrodes. The electrodes are generally at a fixed position relative to the lead tip. Some pacemakers are leadless and have electrodes integrated into the pacemaker housing.

[0006] The pulse generator produces electrical pulses that are transmitted to the heart wall via the leads and electrodes to initiate electrical activity in the heart. A pacemaker may also include a sensing mechanism with sensors that can detect intrinsic or extrinsic electrical activity of the heart. A pacemaker may also include a programmable controller for controlling delivery of the electrical pulses and sensing of electrical activity. The programmable controller has various parameters for controlling the delivery of electrical pulses. Based on the parameter settings, the programmable controller controls the timing, output voltage, and output current of the electrical pulses delivered to each electrode. The timing may be at a predefined rate or responsive to detected electrical activity and may factor in a programmable refractory period. The combination of the lead locations and parameters settings is referred to the pacemaker configuration.

[0007] LBBAP is often employed to treat a patient whose left ventricle is not depolarizing correctly, for example, because of a left bundle branch block (LBBB). LBBAP is a type of cardiac physiologic pacing that involves implanting a lead of an LBBAP pacemaker in the interventricular septum and positioning a right atrial (RA) lead in the RA. When paced, the LBBAP lead provides direct stimulation of the LBB. For patients in sinus rhythm, the RA lead senses intrinsic activity in the RA and signals activation of the LBBAP lead to initiate LV depolarization. These sensed events and stimulations help to improve synchronicity both between the atria and ventricles and between the left and right ventricles, which can result in improved left ventricular ejection fraction. In some circumstances, LBBAP has been shown to be superior to His-bundle pacing with respect to capture thresholds and lead stability, with fewer adverse events.

[0008] An LBBAP procedure begins with the insertion of catheter (with a sheath and a lead with electrodes) through the venous system, typically via the left subclavianor left axillary vein, into the right atrium, and then maneuvered into the right ventricle. The lead is advanced towards the interventricular septum to a position that is typically 1 .5 to 2.0 cm apically from location of the His bundle. When the lead is in contact with the heart wall and correctly positioned, it is advanced (tunneled) into the heart wall. As is it advanced, the lead is paced to determine whether the LBB is “captured.” The determination of whether the LBB is captured is typically based on analysis of the surface electrocardiogram collected while pacing. While advancing the lead and when the LBB is at its implantation site, the parameters of the pacemaker may be varied to determine the various thresholds for the parameters. One threshold is the left bundle branch capture threshold which is the minimum voltage needed to capture the LBB. After the lead is implanted, the RA lead of the pacemaker is implanted into the right atrium for patients without permanent atrial fibrillation.

[0009] Several technologies have been used to help guide the lead to desired position at which to advance the lead into the interventricular septum, but they have disadvantages. A first technology determines the lead location within the body using electrical impedance and / or references to external magnets but requires additional mapping patches to be placed on the patient and / or external magnets to be placed near the patient during a procedure. A second technology uses fluoroscopy to determine the device location but exposes the patient to a contrast agent and the patient and medical personnel to x-rays. The exposure to a contrast agent can cause severe reactions such as pulmonary edema, hypotension, and even cardiac arrest. Fluoroscopy also requires rotating the imaging plane during the LBBAP implantation procedure. A third technology employs intracardiac echocardiography (ICE) imaging to visualize cardiac chambers and the device location but requires manual adjustment of the image plane to perform the mapping process.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 illustrates a baseline ECG of a patient that indicates presence of a Mobitz II AV block.

[0011] Figure 2 illustrates a window generated by the LBBAP system in some embodiments.

[0012] Figure 3 illustrates an updated window generated by the LBBAP system in some embodiments.

[0013] Figure 4 illustrates an ECG indicating that the lead position is at an appropriate implantation site.

[0014] Figure 5 illustrates indicators of an effective position.

[0015] Figure 6 is a flow diagram that illustrates processing of an implantation system in some embodiments.

[0016] Figure 7 is a block diagram that illustrates components of the implantation system in some embodiments.

[0017] Figure 8 is a flow diagram of a controller of the implantation system in some embodiments.

[0018] Figure 9 is a flow diagram that illustrates the processing of a display graphics component of the implantation system in some embodiments.

[0019] Figure 10 is a flow diagram that illustrates the processing of a move to target zone component of the implantation system in some embodiments.

[0020] Figure 11 is a flow diagram that illustrates the processing of a locate component of the implantation system in some embodiments.DETAILED DESCRIPTION

[0021] An implantation system is described that reduces or eliminates the disadvantages associated with current technologies that are employed during an implantation procedure. Such disadvantages may include: complexities associated with use of mapping patches and magnets, adverse reactions to contrast agents; safety concerns related fluoroscopy exposure; complexities and costs associated with ICE imaging; and increased risk of adverse events resulting from prolonged implantation procedure durations.

[0022] The implantation system may be employed to facilitate LBBAP, right bundle branch area pacing (RBBAP), and His bundle pacing (HBP). Although primarily described in the context of an LBBAP system, techniques of the LBBAP system may also be applied in an RBBAP system configured to facilitate RBBAP, and in an HBP system configured to facilitate HBP.

[0023] In some embodiments, the implantation system may be employed as part of a method for treating a patient. Such a method may include: collecting patient information (e.g., electrocardiogram (ECG) and cardiac imaging data); utilizing the implantation system to assist in planning and providing guidance during an implantation procedure; and implanting a pacemaker.

[0024] In some embodiments, the LBBAP system facilitates overcoming these disadvantages and other disadvantages. The LBBAP system may provide an assessment of whether LBBAP would be beneficial for a patient, based on analysis of a baseline ECG, with or without consideration of left ventricular ejection fraction (LVEF) and / or left ventricular activation time (LVAT).

[0025] The LBBAP system may identify a target zone for lead implantation as well as one or more avoidance zones. The LBBAP system may track the position of the lead as it is guided toward the target zone comparing real-time paced QRS morphologies to a library of simulated and / or patient-derived QRS morphologies, each associated with known pacing locations.

[0026] The LBBAP system may determine whether the lead is at an appropriate implantation site (which may be in the target zone) based on an appropriate implantation site (AIS) criterion. The AIS criterion may be determined using mapping results in conjunction with 12-lead ECG features indicative of successful LBBAP implantation outcomes. The AIS criterion may be based on various AIS factors, including:• P-wave morphology for the RA,• QRS morphology for a ventricle,• pacing threshold (e.g., <1 .0 V at 0.5 ms pulse width),• sensing amplitude,• impedance within a normal range (e.g., 300-1200 ohms), and• the implantation site being within the RA appendage.In some embodiments, the AIS criterion may be satisfied based on similarity between of the AIS factors and patient factors. This similarity may be determined through quantitative comparison (e.g., cosine of ECGs of ECG waveforms), range matching of patient factors to AIS factors, or other metrics.

[0027] Alternatively, the AIS criterion may be determined using an AIS machine learning (ML) model trained using a training dataset derived from EHRs. This trainingdataset may include examples with features derived from the AIS factors and labels based on implantation success. This training dataset may also include examples derived from simulations of electrical activity of hearts having various cardiac characteristics, implantation sites, and so on.

[0028] The LBBAP system additionally provides support in determining an appropriate pacemaker configuration and tunneling depth. In some embodiments, the LBBAP system may also track an RA lead as it is guided to an RA implantation site if indicated.

[0029] In some embodiments, to assess whether LBBAP may be beneficial for a patient, the LBBAP system is configured to analyze a baseline ECG in combination with structural and functional cardiac information derived from cardiac imaging. This analysis may be performed using an assessment technique, which may include an assessment ML model or a non-ML assessment algorithm.

[0030] When employing a non-ML assessment algorithm, the LBBAP system may analyze ECG morphology to identify one or more characteristics indicative of the conduction abnormalities such as a left bundle branch block (LBBB) or a Mobitz II atrioventricular (AV) block. For an LBBB assessment, the LBBAP system may factor in features including, but not limited to:• Prolonged QRS duration (e.g., greater than 120 milliseconds),• Broad, notched, or slurred R waves in ECG leads I, aVL, V5, and V6,• Absence of Q waves in leads I, V5, and V6, and• A predominantly negative QRS complex in lead V1 (e.g., deep S wave).The LBBAP system may further determine the severity of the LBBB, including a complete LBBB, incomplete LBBB, a rate-dependent LBBB, and so on.

[0031] For an assessment relating to a Mobitz II AV block, the LBBAP system may factor features such as:• intermittent non-conducted P waves,• constant PR interval in conducted beats, dropped beats, and fixed ratio of P waves.Figure 1 illustrates a baseline ECG of a patient that indicates presence of a Mobitz II AV block. In such a case, when a pacemaker is implanted, the patient may require near-continuous ventricular pacing (e.g., close to 100%).

[0032] The LBBAP system may train an assessment ML model using a training dataset derived from electronic health records (EHRs) of patients prior to an LBBAP procedure. The training dataset includes examples that each comprises a feature vector with one or more features and one or more labels.

[0033] The features may include a voltage-time series for various ECG leads of a baseline ECG and / or characteristics of the voltage-time series such as R waves, QRS complexes, and QRS duration of patients who have had an LBBAP procedure. (Additional features are described in Perez-Riera, A.R., Barbosa-Barros, R., de Rezende Barbosa, M.P., Daminello-Raimundo, R., de Abreu, L.C. and Nikus, K., 2018. Left bundle branch block: Epidemiology, etiology, anatomic features, electrovectorcardiography, and classification proposal. Annals of Noninvasive Electrocardiology: The Official Journal of the International Society for Holter and Noninvasive Electrocardiology, Inc, 24(2), p.e12572, which is hereby incorporated by reference.) A feature vector may also include features derived from the cardiac morphology of the patient such as heart wall thicknesses, LVEF, and LVAT.

[0034] A label for an example may indicate whether the feature vector represents an LBBB or Mobitz II AV block and, if so, an assessment (e.g., probability) of whether LBBAP was beneficial to the patient which may be derived from a post-procedure ECG. The label may also indicate the implantation site.

[0035] The assessment ML model may employ various ML architectures such as a neural network (NN), a convolutional neural network (CNN), K-means clustering, a decision tree, and so on. To provide an assessment for a patient, the LBBAP system generates a feature vector based on a baseline ECG of the patient and optionally cardiac morphology and applies the assessment ML model to the feature vector to generate the assessment.

[0036] The LBBAP system may analyze the cardiac morphology of the patient’s heart. The analysis may be based on images of the heart such as computed tomography (CT) images, magnetic resonance imaging (MRI) images, echocardiogram images, and so on. From the analysis of the images, the LBBAP system may determinevarious characteristics of the cardiac morphology such as LVEF, LV geometry (dimension and volume), LV wall motion and strain, LV wall thickness, mitral valve function, and so on which may be features employed when training the assessment ML model. Some of these characteristics may affect the assessment whether a location is an appropriate implantation site. For example, the LV wall motion may indicate regions with impaired contractility which may tend to indicate that a location is not an appropriate implantation site. The LBBAP system may also base the assessment on other factors such as a New York Heart Association (NYHA) classification, previous responses to pacing or CRT, and presence of scar tissue or fibrosis based upon intracardiac voltage mapping or late gadolinium enhancement on cardiac MRI. Techniques for determining various characteristics of a heart (e.g., LV wall motion) are described in PCT App. No. PCT / US25 / 19480 titled “3D Cardiac Visualization System” and filed on March 12, 2025, which is hereby incorporated by reference.

[0037] To identify the target zone and the avoidance zone, the LBBAP system may employ a generic heart model or a patient-specific heart model. The target zone may have a diameter of 0.5 cm that is centered 1.5 to 2.0 cm apically from the His bundle location as indicated on the heart model. The avoidance zone may be centered on the right ventricular outflow tract, which may be identified, for example, based on a segmentation of a 3D graphic of a heart. Techniques for generating a segmentation are described A technique for segmentation is described in Kong, F., Wilson, N. and Shadden, S., 2021 . A deep-learning approach for direct whole-heart mesh reconstruction. Medical image analysis, 74, p.102222 and described in Ronneberger, O., Fischer, P. and Brox, T., 2015. U-net: Convolutional networks for biomedical image segmentation. In Medical image computing and computer-assisted intervention- MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18 (pp. 234-241 ). Springer International Publishing, which are hereby incorporated by reference. Rather than having a fixed location and size for a target zone, the LBBAP system may identify a target zone based on analysis of EHRs of patients who have had LBBAP procedures. For example, a pediatric patient may have a heart which is smaller than the standard adult model. In such a patient, the target zone may be proportionally closer to the tricuspid valve annulus and proportionally smaller than the baseline target zone. Alternatively, patients with different forms of adult congenital heart disease (ACHD) may have variations in the location oftheir cardiac conduction system, which varies by the type of ACHD and severity. The target zone may be adjusted based upon patient-specific cardiac imaging or based upon machine-learning analysis of similar patients.

[0038] A target zone ML model inputs a feature vector with some features that are derived from cardiac morphology and a baseline ECG of a patient and outputs an indication of a target zone. The target zone ML model may be trained using feature vectors and labels derived from EHRs of patients who have had an LBBAP procedure. The feature vector derived from the EHR of a patient may include features derived from a cardiac image and a baseline ECG of the patient. The label for a feature vector is the target zone.

[0039] The target zones used in training the target zone ML model may be centered on the implantation site for the patient. The area of the target zone may be fixed or variable. To determine a variable area for a patient, the LBBAP system may identify clusters of patients with similar cardiac morphologies and baseline ECGs. For each cluster, the LBBAP system may calculate a mean implantation site for the patients in the cluster and a standard deviation assuming, for example, a Gaussian distribution. The LBBAP system may then assign a target zone to each cluster that is centered on the mean implantation site and that has an area (or more generally shape) that is based on the implantation sites within, for example, one standard deviation.

[0040] The training dataset for the target zone ML model may include examples that each comprise a feature vector of patients labeled with the target zone of the cluster that the patient is within. Such a target zone ML model may be, for example, an NN or a CNN. The LBBAP system may employ K-means clustering to generate the clusters. The LBBAP system may alternatively identify the target zone for a patient based on similarity of the patient’s cardiac morphology and baseline ECG to the means of the clusters. The similarity to the means may be assessed based on a similarity score (e.g., cosine similarity) with similar being considered to be above a threshold similarity score (e.g., 0.95) or the highest similarity score.

[0041] In some embodiment, the LBBAP system may generate training data sets for the various ML models using simulated data. The LBBAP system run simulations of electrical activity of a heart assuming various cardiac morphologies and various types of conduction blocks. Each simulation is run until the electrical activity stabilizes.Techniques for assessing such stability are described in Krummen, D.E., Hayase, J., Morris, D.J., Ho, J., Smetak, M.R., Clopton, P., Rappel, W.J. and Narayan, S.M., 2014. Rotor stability separates sustained ventricular fibrillation from self-terminating episodes in humans. Journal of the American College of Cardiology, 63(24), pp.2712-2721 , which is hereby incorporated by reference. For each simulation, the LBBAP system may run multiple simulations assuming different pacing location of the IV septum. The LBBAP system then generates a simulated ECG based on the simulated electrical activity of each simulation. If a simulated ECG represents an ECG that is similar to a sinus rhythm ECG, the pacing position is indicated as being at an effective position for implantation. The similarity of ECGs may be determined based on a Pearson’s correlation, Euclidian distance, cosine similarity, and so on. The data of the simulations may be used as the training dataset for an AIS model and an effective position ML model, which are described below. Given an effective position of a simulation, an appropriate implantation site may be, for example, the location on the RV-side of the IV septum that is closest to the effective position. Techniques for simulating electrical activity of a heart are described in U.S. Pat. No. 10,856,812 titled “Machine Learning using Simulated Cardiograms” and issued on December 8, 2020, which is hereby incorporated by reference.

[0042] Figure 2 illustrates a window generated by the LBBAP system in some embodiments. The window 200 includes an ECG area 210, a heart graphic area 220, and a beat information area 230. The ECG area displays in real time ECGs collected from the patient that may be a baseline ECG or a pacing ECG. Although not illustrated, the ECG area may provide tools for manual selection of portions of an ECG to be used by the LBBAP system to, for example, identify the location of the lead. The heart graphic area includes a graphic of a heart 221 , a lead position area 222, a target zone area 223, and an avoidance zone area 224. The graphic may be based on a generic heart model or a patient-specific heart model. The lead position area as described below indicates the current position of the lead. The target zone area indicates the target zone, and the avoidance zone area indicates the avoidance zone. The beat information area allows a user to select beats to be used when performing various analyses such as determining an appropriate implantation site. Figure 3 illustrates an updated window generated by the LBBAP system in some embodiments. The lead position area 322 is shown closer to the target zone 323 than in Figure 2. To move to this lead position area, the leadwas pulled back a small amount toward the tricuspid valve. In some embodiments, the LBBAP system may superimpose the target zone and the avoidance zone on a fluoroscopy image collected during the procedure. As described below, the LBBAP system may also superimpose a candidate path that provides an indication of a path to the target zone.

[0043] The patient-specific heart model may also be generated based on heart model mappings of cardiograms (e.g., ECGs or vectorcardiogram (VCGs)) to heart models employed in simulations of electromechanical activity of a heart or derived from clinical data. Techniques for running such simulations are described in U.S. Pat. No. 10,856,812 titled “Machine Learning using Simulated Cardiograms” and issued on December s, 2020, which is hereby incorporated by reference. The ‘812 patent describes source location mappings of simulated cardiograms to source locations of arrhythmia. Each source location mapping is generated from a simulation based on the source location employed in the simulation and a simulated cardiogram generated from a simulation. Since each simulation is based on a heart model, each heart model mapping may be generated by mapping the simulated cardiogram of a simulation to the heart model of that simulation. The clinical data of EHRs may include three-dimensional (3D) images of a heart and cardiograms collected when the 3D images were collected. The techniques described in PCT App. No. PCT / US25 / 19480 and described in Int. Pub. No. 2023 / 156017 titled “Overall Ablation Workflow System” and published on September 7, 2020, which are hereby incorporated by reference, may be employed to generate the heart models from the 3D images. A heart model mapping maps a clinical cardiogram of an EHR to the heart model generated from a 3D image of that EHR. A generic heart model may be selected (e.g., by an electrophysiologist) from a predefined set of generic heart models based on what may be closest to the geometry of the patient’s heart.

[0044] During implantation, the LBBAP system continually identifies the lead location of the lead tip as it is moves through the right ventricle to the implantation site based upon the paced QRS morphology. In some embodiments, the LBBAP system employs techniques that do not require alternative imaging data such as images collected using fluoroscopy. As a result, the LBBAP system minimizes the patient’s exposure and risks associated with use of fluoroscopy with or without use of an intravenous contrast agent. When the lead is paced, the LBBAP system identifies thelead position based upon the paced QRS morphology and displays an indication of the lead position using a lead position area indicator 222. In this way, the current lead position is shown relative to the target zone and the avoidance zone to assist in guiding the lead to the target zone. Techniques for identifying the lead position based on an ECG collected while pacing the leads and superimposing a candidate path are described in U.S. Pat. No. 1 1 ,338,131 titled “Guiding Implantation of an Energy Delivery Component in a Body,” and issued on May 24, 2022, which is hereby incorporated by reference. Additional techniques for determining the lead location when pacing does not capture cardiac tissue are described in PCT App. No. US24 / 16434 titled “Stimulation Device Location Identification System” and February 12, 2024, which is hereby incorporated by reference.

[0045] The ECG areas 210 and 310 display an ECG as it is collected from the patient during the LBBAP procedure. When the lead position is within the target zone, the LBBAP system employs an AIS ML model or a non-ML algorithm to assess whether the lead position is at an appropriate implantation site. The assessment may be based on the morphologies of ECG leads II, III, aVR, aVL, and V1. The morphologies indicating that the lead position is at an appropriate implantation site may include (1 ) lead II with a QRS amplitude that is greater than lead III, (2) lead aVR and lead aVL QRS complexes being discordant, and (3) notching in the negative portion of the QRS complex in lead V1. The morphologies may be considered sub-criteria of an implantation site criterion. Figure 4 illustrates an ECG indicating that the lead position is at an appropriate implantation site. The ECG leads employed in the assessment are indicated by the circles. The circles may be initially yellow to indicate that an ECG lead does not have a morphology indicating an appropriate implantation site. When an ECG lead does have such a morphology, the circle is switched to be green. Some morphologies indicating that the lead may be at an appropriate implantation site for HBP are described in Vijayaraman, P., Chung, M.K., Dandamudi, G., Upadhyay, G.A., Krishnan, K., Crossley, G., Bova Campbell, K., Lee, B.K., Refaat, M.M., Saksena, S. and Fisher, J.D., 2018. His bundle pacing. Journal of the American College of Cardiology, 72(8), pp.927-947, which is hereby incorporated by reference. Some morphologies indicating that the lead may be at an appropriate implantation site for LBBAP, RBBAP, and HBP are described in Jastrz^bski, M., 2021. ECG and pacingcriteria for differentiating conduction system pacing from myocardial pacing. Arrhythmia & electrophysiology review, 10(3), p.172, which is hereby incorporated by reference.

[0046] The AIS ML model may be trained with feature vectors and labels of examples of a training data set. A feature vector may include various features such as voltage-time series representing the ECG leads II, III, aVR, aVL, and V1 or characteristics derived from these ECG leads such as QRS amplitude. A label indicates whether or not the features represent an appropriate implantation site. The appropriate implantation site model may have ML sub-models that are trained to separately identify the three morphologies described above. While a morphology is satisfied, the LBBAP system displays the circle for the ECG leads associated with that morphology in green. In addition to providing green as a cue, the LBBAP system may also provide various audible cues to indicate, for example, that the lead is in the target zone or the avoidance zone or at an AIS.

[0047] When the lead is at an appropriate implantation site, the electrophysiologist may start tunneling the lead into the IV septum. The LBBAP system provides an assessment of whether the lead is at an effective position within the IV septum to ensure effective capture of the LBB which may be at a depth of 1 .0 to 1 .5 cm. The effectiveness of the lead positions during tunneling may be based on one or more of ECG morphology, LBBAP lead sensing, pacing impedance, and pacing thresholds. The lead position is considered effective when an effectiveness criterion is satisfied based on one or more sub-criteria. The sub-criteria may be based on a terminal positive wave in lead V1 and a short (<85-90 ms) stimulus to peak QRS duration in lead V5 (in this example 61 ms). Figure 5 illustrates indicators of an effective position. The stars may be initially displayed in yellow and switch to green when the lead V1 or V5 portion of the effectiveness criterion is satisfied. The LBBAP system may also provide an audio cue when one or both of the portions have the appropriate morphology. When the lead position is an effective position, implantation of the lead is complete, and the sheath employed in the implantation can be removed. The lead can then be connected to the pulse generator of the pacemaker.

[0048] For patients undergoing a right atrial (RA) lead implantation, the ECG localization process described above may be used to guide placement of the RA lead in an appropriate location — such as the right atrial appendage, the lateral wall of theright atrium, or another site deemed suitable by the attending electrophysiologist — based on the paced P-wave morphology compared to both the simulation and clinical libraries used for mapping.

[0049] In some embodiments, the LBBAP system may identify a candidate path for movement of the lead to the target zone. The LBBAP system may employ a path planning (PP) system to generate the candidate path. A PP system may employ ML techniques and / or non-ML algorithmic techniques to identify candidate paths to help inform movement of the lead to the target zone. The PP system identifies candidate paths based on PP input data that includes patient data (e.g., cardiac characteristics), planning data (e.g., target zone), and / or a catheter specification (e.g., flexibility). When employing an ML technique, the PP system generates a feature vector with one or more features based on the PP input data and applies a PP ML model to the feature vector to generate a candidate path. The PP ML model may be trained using PP input data extracted from EHRs of patients who have had successful cardiac procedures. When employing a non-ML algorithmic technique, the PP system may employ a PP algorithm that searches a space of possible paths to identify one or more candidate paths factoring in the PP input data. The PP algorithm may employ various techniques such as a breadth-first search or a depth-first search of the space of possible paths. The candidate paths identified via the PP algorithm may be employed to train a PP ML model instead of or in addition to the EHR data or may be employed to bootstrap the training of the PP ML model. A candidate path may be provided to a catheter navigation device that may employ robotic magnetic navigation to assist in navigation of a catheter. A candidate path may also be displayed superimposed on the graphic 221 that includes the lead position, the target zone, and the avoidance zone(s). The PP system may also be employed during a procedure to generate a new candidate path given the current location of the lead, for example, factoring new patient data. The PP system provides real-time, data-driven recommendations for path planning. As a result, the PP system tends to generate improved candidate paths that may lead to reduced procedure times and improved patient outcomes. Techniques for path planning are described in PCT App. No. US25 / 14448 titled “Cardiac Catheter Path Planning System” and filed on February 4, 2025, which is hereby incorporated by reference.

[0050] In some embodiments, a person may manually demarcate portions of an ECG that are to be used by the LBBAP system to perform various analyses such as anassessment of the benefit of LBBAP or a determination of whether a location is an appropriate implantation site or an effective position. A demarcated portion ideally would encompass only the portion that is most relevant to the determination. However, it can be difficult for a person to demarcate only the most relevant portion. To account for this difficulty, the LBBAP system may refine a manually demarcated portion so that the portion more accurately represents the most relevant portion. Techniques for manually demarcating a portion and refining a portion are described in Int. Pub. No. WO 2024 / 044719 titled “Automatic Refinement of an Electrogram Selection” and published on February 29, 2024, which is hereby incorporated by reference. The ‘719 publication also describes techniques for automatically (rather than manually) demarcating a portion of an ECG. For example, to identify an R wave, the LBBAP system may select the highest peak of a cardiac cycle as representing an R peak. The LBBAP system may then identify the start and end of the R wave using a gradient descent search on the voltage-time series to identify valleys to the left and right of the R peak.

[0051] Figure 6 is a flow diagram that illustrates processing of an implantation system in some embodiments. The implantation system may be, for example, an LBBAP system, an RBBAP system, or an HBP system. The implantation system 600 displays a heart graphic along with lead locations, identifies an appropriate implantation site, and identifies an effective position along with a suitable configuration for a pacemaker. In block 601 , a component of the implantation system displays a heart graphic that includes indications of the target zone and an avoidance zone. In blocks 602-604, the component identifies an appropriate implantation site. In block 602, the component receives a location ECG. In block 603, the component identifies the current location of the lead (e.g., when the lead tip of the lead tip captures myocardial tissue during pacing). The component also displays an indication of the current location on the heart graphic. In decision block 604, if an appropriate implantation site (AIS) criterion has been satisfied, then the component continues at block 605, else the component loops to block 602 to receive the next location ECG. In block 605, the component displays an indication that an appropriate implantation site has been identified. In blocks 606-607, the component identifies an effective position (EP) for the lead. In block 606, the component receives a tunneling ECG. In decision block 607, if an effective position criterion is satisfied, then the component continues at block 608, else the component loops to block 606 to receive the next tunneling ECG. In block 608,the component displays an indication that an effective position has been identified and then completes. The component may identify and display an indicator of a pacemaker configuration that may be appropriate. The identification of an appropriate pacemaker configuration may be based on a configuration ML model that is trained using training dataset derived from EHRs of patients who had LBBAPs. The examples of the training dataset includes feature vectors of feature may include ECGs, cardiac morphology, effective position, and so on labeled with the pacemaker configuration associated with the effective position. To identify a pacemaker configuration for LBBAP, the features derived from EHR of the patient and the current position during tunneling are input to configuration ML model which outputs a pacemaker configuration.

[0052] Figure 7 is a block diagram that illustrates components of the implantation system in some embodiments. The implantation system 700 includes a controller 701 , a move to target zone (TZ) component 702, an identify AIS component 703, an identify effective position component 704, a display graphics component 705, a target zone ML model 706, an AIS ML model 707, an effective position ML model 708, an update graphics component 709, a receive ECG component 710, an identify path component 71 1 , and a control navigation device component 712. The implantation system includes an interface to a mapping system 720, an ECG collection device 730, an EHR data store 740, an ML weights data store 750, and a navigation device 760. The controller controls the overall processing of the implantation system. The move to target zone component identifies the current location of the lead based on an ECG and provides an indication of the current location. The identify AIS component identifies the current location of the lead based on an ECG and determines whether the current lead location is an appropriate implantation site based on an appropriate implantation site criterion. The identify effective position component identifies while tunneling whether the position of the lead with the current pacemaker configuration is an effective position based on an effective position criterion. The display graphics component displays a heart graphic that indicates a target zone and an avoidance zone. The target zone ML model inputs various features of the patient and outputs a target zone. The AIS ML model inputs an ECG and outputs an indication of whether the current location is an AIS. The effective position ML model inputs an ECG and outputs whether the lead is at an effective position with the current pacemaker configuration. The effective position ML model may be trained with a training dataset of examples derived from simulation and / or EHRs.The update graphics component dynamically updates the heart graphic when an ECG is received that was collected while pacing. The update graphic component displays the ECG, an indication of the current location, and an indication of whether the various criteria have been satisfied. The receive ECG component receives ECGs from the ECG collection device. The identify path component identifies a path from the current location of the lead to the target zone. The control navigation device component identifies a next location along the path and directs the navigation device to move the lead to that location. The current location of the lead is identified by inputting an ECG to the mapping system which outputs a location as described, for example, in the ‘812 patent.

[0053] Figure 8 is a flow diagram of a controller of the implantation system in some embodiments. The controller 800 assists in guiding the lead to an effective position. In block 801 , the controller invokes a display graphics component to display a heart graphic along with an indication of the target zone and the avoidance zone. In block802, the component invokes the move to target zone component to monitor movement of the lead and indicate when the lead is at a location within the target zone. In block803, the component invokes the AIS component to determine whether the lead is at an appropriate implantation site based on satisfying an AIS criterion. In block 804, the component invokes a locate effective position component to identify whether the lead is at an effective position. The component then completes.

[0054] Figure 9 is a flow diagram that illustrates the processing of a display graphics component of the implantation system in some embodiments. The display graphic component 900 generates a heart graphic based on imaging data of the heart, identifies a target zone and an avoidance zone, adds indications of the target zone and the avoidance zone to the heart graphic, and then outputs the heart graphic. In block 901 , the component retrieves imaging data of a heart from an EHR of the patient. In block 902, the component generates a segmentation of the imaging data. In block 903, the component generates a 3D mesh of the heart based on the segmentation. In block 904, the component generates a heart graphic based on the 3D mesh. In block 905, the component applies a target zone ML model that inputs the 3D mesh and outputs an indication of a target zone. In block 906, the component identifies an avoidance zone which may correspond to a standard location within the heart. In block 907, thecomponent adds an indication of the target zone and the avoidance zone to the heart graphic. In block 908, the component displays the heart graphic and then completes.

[0055] In some embodiments, the graphic of the heart may be generated from a 3D image (e.g., CT scan) of the patient’s heart. The implantation system may employ the Blender open-source system to generate and the heart graphic. (Blender Foundation, Blender, version 2.93, Blender, 2023. [Online]. Available: https: / / www.blender.org / .) Various techniques for generating a 3D graphic and for augmenting the 3D graphic with a location derived from an ECG are described in Int. Pub. No. WO 2023 / 0168017 titled “Overall Ablation Workflow System” and published on September 7, 2023, which is hereby incorporated by reference.

[0056] Figure 10 is a flow diagram that illustrates the processing of a move to target zone component of the implantation system in some embodiments. The move the target zone component 1000 is invoked to track the location of the lead and displays indications of the locations. In block 1001 , the component employs the path planning system to identify a path from the current position of the lead to the target zone. The path planning system may factor in the cardiac geometry of the patient (e.g., derived from imaging data) and characteristics of the catheter (e.g., flexibility). In block 1002, the component displays an indication of the path on the heart graphic. In block 1003, the component receives a location ECG from the ECG collection device. In block 1004, the component provides the location ECG to the mapping system and receives a location from the mapping system. In block 1005, the component displays an indication of the location on the heart graphic. In decision block 1006, if the location is within the target zone, then the component continues at block 1008, else the component loops to block 1007. In decision block 1007, if the location is on the identified path, then the component continues at block 1003 to receive the next location ECG, else the component loops to block 1001 to identify a new path to the target zone given the current location. In block 1008, the component outputs an indication that the current location is within the target zone and then completes.

[0057] Figure 11 is a flow diagram that illustrates the processing of a locate component of the implantation system in some embodiments. The locate component 1 100 is invoked to identify when the lead is at an AIS or an effective position. The overall processing of the identification of whether a lead is an AIS or an effective positionis similar except that different criterion are employed and different ML models or non- ML algorithms are employed. In block 1101 , the component collects an ECG from the ECG collection device. In blocks 1 102-1104, the component loops determining whether each sub-criterion of the AIS criterion or the effective position criterion is satisfied. In block 1 102, the component selects the next sub-criterion. In decision block 1 103, if all the sub-criterion have already been selected, then the component continues at block 1105, else the component continues at block 1104. In block 1 104, the component evaluates the sub-criterion to determine whether it is satisfied and then loops to block 1102 to select the next sub-criterion. In block 1105, the component displays an indication of whether each sub-criterion has been satisfied. In decision block 1106, if the criterion has been satisfied, then the component completes, else the component loops to block 1101 to select the next ECG.

[0058] The term “lead” has different meanings in cardiology depending on context. For example, an ECG comprises “leads” that each is a voltage-time series (e.g., leads I, aVL, and V5) that is often displayed as a graph. In contrast, a pacemaker device has a “lead” that is a physical wire connecting a pulse generator to an electrode at the lead tip. The meaning of “lead” as used in this detailed description is clear based on the context in which it is used.

[0059] The computing systems (e.g., network nodes or collections of network nodes) on which the implantation system and the other described systems may be implemented may include a central processing unit, input devices, output devices (e.g., display devices and speakers), storage devices (e.g., memory and disk drives), network interfaces, graphics processing units, communications links (e.g., Ethernet, Wi-Fi, cellular, and Bluetooth), and so on. The input devices may include keyboards, pointing devices, touch screens, gesture recognition devices (e.g., for air gestures), head and eye tracking devices, microphones for voice recognition, and so on. The computing systems may include high-performance computing systems, distributed systems, cloudbased computing systems, client computing systems that interact with cloud-based computing system, desktop computers, laptops, tablets, smartphones, gaming devices, servers, and so on. The computing systems may access computer-readable media that include computer-readable storage mediums and data transmission mediums. The computer-readable storage mediums are tangible storage means that do not include a transitory, propagating signal. Examples of computer-readable storage mediumsinclude memory such as primary memory, cache memory, and secondary memory (e.g., DVD), and other storage. The computer-readable storage media may have recorded on them or may be encoded with computer-executable instructions or logic that implements the implantation system and the other described systems. The data transmission media are used for transmitting data via transitory, propagating signals or carrier waves (e.g., electromagnetism) via a wired or wireless connection. The computing systems may include a secure crypto processor as part of a central processing unit (e.g., Intel Secure Guard Extension (SGX)) for generating and securely storing keys and for encrypting and decrypting data using the keys and for securely executing all or some of the computer-executable instructions of the PCI system. Some of the data sent by and received by the implantation system may be encrypted, for example, to preserve patient privacy (e.g., to comply with government regulations such the European General Data Protection Regulation (GDPR) or the Health Insurance Portability and Accountability Act (HIPAA) of the United States). The implantation system may employ asymmetric encryption (e.g., using private and public keys of the Rivest-Shamir-Adleman (RSA) standard) or symmetric encryption (e.g., using a symmetric key of the Advanced Encryption Standard (AES)).

[0060] The one or more computing systems may include client-side computing systems and cloud-based computing systems (e.g., public or private) that each executes computer-executable instructions of the implantation system. A client-side computing system may send data to and receive data from one or more servers of the cloud-based computing systems of one or more cloud data centers. For example, a client-side computing system may send a request to a cloud-based computing system that hosts the mapping system or provides an ECG collection interface. A cloud-based computing system may respond to the request by sending to the client-side computing system data derived from performing task such as identifying a location given an ECG. The servers may perform computationally expensive tasks in advance of processing by a client-side computing system such as training a ML model in response to data received from a client-side computing system. A client-side computing system may provide a user experience (e.g., user interface illustrated in Figures 2-5) to a user of the implantation system. The user experience may originate from a client computing device or a server computing device. For example, a client computing device may generate a patient-specific graphic of a heart and display the graphic. Alternatively, a cloud-basedcomputing system may generate the graphic (e.g., in a Hyper-Text Markup Language (HTML) format, an extensible Markup Language (XML) format or a Digital Imaging and Communications in Medicine (DICOM) format) and provide it to the client-side computing system for display. A client-side computing system may also send data to and receive data from various medical devices such as an ECG collection device or an imaging device. The data received from the medical devices may include an ECG and a computed tomography (CT scan). The data sent to a medical device may include data, for example, in the DICOM format. A client-side computing device may also send data to and receive data from medical computing systems. Such medical computing systems store patient medical history data (EHRs systems), descriptions of medical devices (e.g., type, manufacturer, and model number of a catheter) of a medical facility, results of procedures, and so on. The term cloud-based computing system may encompass computing systems of a public cloud data center provided by a cloud provider (e.g., Azure provided by Microsoft Corporation) or computing systems of a private server farm (e.g., operated by the provider of the implantation system).

[0061] The implantation system and the other described systems are implemented using computer-executable instructions, such as program modules and components, executed by one or more computers, processors, or other devices. Generally, program modules or components include routines, programs, objects, data structures, and so on that perform tasks or implement data types of the implantation system and the other described systems. Typically, the functionality of the program modules may be combined or distributed as desired in various examples. Aspects of the implantation system and the other described systems may be implemented in hardware using, for example, an application-specific integrated circuit (ASIC), graphic processing units (GPU), or a field programmable gate array (FPGA). For example, a GPU may be employed when training an ML model.

[0062] An ML model employed by the implantation system may be any of a variety or combination of supervised, semi-supervised, self-supervised, unsupervised, or reinforcement learning ML models including neural network (NN) such as a fully connected, convolutional, recurrent, or autoencoder neural network, a restricted Boltzmann machine, a support vector machine, a Bayesian classifier, K-means clustering, K Nearest Neighbors (KNN), transformer, recommender, and so on.

[0063] When the ML model is a deep neural network, the model is trained using a training dataset that includes examples that each having a feature vector with features derived from data and having labels corresponding to the features. For example, the examples may include feature vectors with features derived from images of ECGs, a voltage-time series representation of an ECGs, or a pacemaker configuration. A feature may be an image of an ECG or the voltage-time series or may be derived from the image or the voltage-time series (e.g., QRS width) and a pacemaker configuration. The labels may be indications of whether the feature vectors represent an effective position.

[0064] The training results in a set of weights for the activation functions of the layers of the deep neural network that are stored in the ML weights data store. The trained deep neural network can then be applied to new data to generate a label for that new data. An ML model may generate values of discreet domain (e.g., location), probabilities, and / or values of a continuous domain (e.g., probability of a location being an appropriate implantation site.

[0065] A NN model has three primary components: architecture, loss function, and search algorithm. The architecture defines the functional form relating the inputs to the outputs (in terms of network topology, unit connectivity, and activation functions).

[0066] The training involves searching for a set of weights that minimizes the loss function. In some embodiments, the NN model may employ a radial basis function (RBF) and utilize a standard or stochastic gradient descent as the optimization technique, optionally incorporating backpropagation.

[0067] A NN has multiple layers such as a convolutional layer of a convolutional neural network (CNN), a rectified linear unit (ReLU) layer, a pooling layer, a fully connected (FC) layer, and so on. More complex NN architectures may include multiple convolutional layers, ReLU layers, pooling layers, and FC layers. Each layer comprises a plurality of neurons, each configured to generate an output based on the outputs of preceding layers or the original input. A neuron receives one or more inputs and applies an activation function to compute an output value.

[0068] For example, an NN may have multiple ReLU layers that each comprises multiple neurons ( e.g., 10 layers with 128 neurons per layer) that each has an ReLU as an activation function. For example, the architecture of the target zone ML modelmay have an input layer with neurons that each inputs multiple features derived from a segmentation of a CT scan of a heart.

[0069] A convolutional layer may include multiple filters (also referred to as kernels or activation functions). A filter is configured to process a convolutional window, for example, of a CT slice of a heart image, apply a corresponding set of weights to the elements of the window, and generate an output value. For example, if a CT slice is 256 by 256 pixels, the convolutional window may be 8 by 8 pixels, and the filter may apply a distinct set of weights to each of the 64 pixels to compute an output.

[0070] The convolutional layer may include, for each filter, a node (also referred to as a neuron) for each pixel of the CT slice, assuming a stride of one and appropriate padding. Each node outputs a feature value based on the application of the filter’s learned weights. The implantation system may employ a CNN when segmenting a CT scan of a patient.

[0071] An activation function generates a weighted combination of its inputs to produce an output. In some embodiments, the activation function is a ReLU which computes the weight sum of inputs and returns the maximum of zero and the resulting value, i.e., max(0, resulting value), thereby producing non-negative outputs. The weights of the activation functions are learned when training an ML model. The ReLU function may be implemented in a dedicated ReLU layer comprising neurons that receive the outputs of a prior layer and apply the ReLU function to generate corresponding rectified outputs. The weights associated with the activation functions are learned during the training of the ML model.

[0072] A pooling layer may be employed to reduce the spatial dimensions of the preceding layer’s output by performing a downsampling operation. In one example, each neuron in a pooling layer may receive 16 inputs from the prior layer and produce a single output, resulting in a 16-to-1 reduction in dimensionality.

[0073] An FC layer comprises a plurality of neurons, each configured to receive all the outputs from the proceeding layer and computer a weighted combination of those inputs. For instance, if the penultimate layer outputs 256 values and the FC layer has three output neurons corresponding to three classifications, each neuron receives all 256 input values, applies a distinct set of weights, and computes a score associated with its respective classification.

[0074] The LBBAP system may employ k-means clustering to generate clusters of patients with similar cardiac morphologies and baseline ECGs. Given feature vectors representing the training data, k-means clustering clusters the feature vectors into cluster of similar feature vectors. With k-means clustering, the number of clusters may be predefined. For example, the classification system may employ 20 clusters (k=20) to represent 20 different types of cardiac morphologies and baseline ECGs. An example training technique initially randomly places a feature vector in each cluster. The training then repeatedly calculates a mean feature vector of each cluster, selects a feature vector not in a cluster, identifies the cluster whose mean is most similar, adds the feature vector to that cluster, and moves the feature vectors already in the clusters to the cluster with the most similar mean. Similarity may be determined, for example, based on Pearson similarity, cosine similarity, and so on. The training ends when all the feature vectors have been added to a cluster. To determine a target zone, an appropriate implantation site, or effective position for a patient, a feature vector is generated for the patient. The cluster with a mean that is most similar to that feature vector is identified. The target zone, the appropriate implantation site, or effective position associated with that cluster is selected for that patient.

[0075] The LBBAP system may employ an ML decision tree to identify information (e.g., assessment of the benefit of an LBBAP) for a patient. An ML decision may have the same form as a manually generated decision tree. However, the feature associated with each decision node of an ML decision tree may be selected automatically based on analysis of a training data set. Each decision node (i.e., non-leaf node) of a decision tree corresponds to a feature and each branch from a decision node may correspond to a value or range of values for the feature of that decision node. For example, a decision node corresponding LVEF may have branches for low, normal, high, and very high, and a decision node corresponding to LVAT may have slow, normal, or fast. The leaf nodes of the decision tree may indicate assessments of the benefit of an LBBAP. The assessment of a leaf node (assessment node) is intended for a patient with features that match the values of the features along the path from the root node to that leaf node. The LBBAP system may also employ an ML decision tree to identify a target zone, an appropriate implantation site, and an effective position. The LBBAP system may also employ an ML decision tree to determine if the location is an appropriate implantation site or a position is an effective position.

[0076] For an ML decision tree, an entropy score may be used by an ML decision tree generator to select the feature to be associated with each decision node. The entropy score for a possible feature for a decision node is based on the distribution of its values in node feature vectors for that node. A node feature vector for a decision node has the values of the branches along the path from the root node to that decision node. If a first possible feature for a decision node has an equal number of node feature vectors for each value, the entropy for the first possible feature is considered to be high. In contrast, if a second possible feature has node feature vectors for which 75% have the same value, then the entropy for the second possible feature is considered to be low. In such a case, the ML decision tree generator would select the second possible feature for that decision node. The ML decision tree generator may also analyze features with continuous values to identify cut points that tend to minimize entropy. Techniques for identifying cut points are described in Fayyad, U.M. and Irani, K.B., 1992. On the handling in decision tree of continuous-valued attributes generation. Machine Learning, 8, pp.87-102, which is hereby incorporated by reference.

[0077] An ML decision tree generator may employ a depth-first, recursive algorithm to build the ML decision tree. The ML decision tree generator may employ a path termination criterion to determine when to terminate a path. The path termination criterion may be, for example, when the percentage of node feature vectors that are associated with the same assessment is above a threshold percentage. For example, if 80% of the node feature vectors have an assessment that an LBBAP would be beneficial, the ML decision tree generator may add a leaf node that indicates that in 80% of cohort patients the LBBAP was beneficial. Other termination criteria may be that the number of node feature vectors is below a threshold number, or the path has reached a maximum depth. In such cases, the ML decision tree generator may add a leaf node that indicates that an assessment cannot be provided or that the assessment has a low confidence.

[0078] To provide an assessment for a patient, a feature vector is generated for the patient. The values of the features of the patient feature vector are used to identify the path that the patient feature vector matches. The assessment is based on the leaf node of that path.

[0079] The following paragraphs describe various aspects of the implantation system. An implementation of the implantation system may employ any combination or sub-combination of the aspects and may employ additional aspects. The processing of the aspects may be performed by one or more computing systems with one or more processors that execute computer-executable instructions that implement the aspects and that are stored on one or more computer-readable storage mediums.

[0080] In some aspects, the techniques described herein relate to a method performed by one or more computing systems for guiding implantation of a lead of a pacemaker for physiological pacing of a heart of a patient, the method including: displaying a graphical representation of a heart that includes an indication of a target zone and an avoidance zone for implantation of the lead; for each of a plurality of location electrocardiograms collected while pacing the lead at locations within the heart, identifying a location of the lead based on the electrocardiogram; displaying on the graphical representation an indication of identified location; and determining whether the identified location corresponds to an appropriate implantation site based on the electrocardiogram; when the identified location is an appropriate implantation site, outputting an indication that the lead is at an appropriate implantation site; and for each of one or more of tunneling electrocardiograms collected during tunneling of the lead within the heart wall, determining whether the lead is at an effective position within the heart wall based on the tunneling electrocardiograms; and when the lead is at an effective position, outputting an indication that the lead is at an effective position.

[0081] In some aspects, the techniques described herein relate to a method wherein the determination of whether the identified location is an appropriate implantation site is based on an appropriate implantation site criterion that includes subcriteria that is each based on morphology of a location electrocardiogram and further including, for each sub-criterion, displaying a voltage-time series representation of the location electrocardiogram for that sub-criterion with an indication of whether the subcriterion is satisfied.

[0082] In some aspects, the techniques described herein relate to a method wherein the determining of whether the lead is at an effective position is based on an effective location criterion that includes sub-criteria that is each based on morphology of a tunneling electrocardiogram and further including, for each sub-criterion, displayinga voltage-time series representation of the tunneling electrocardiogram for that subcriterion with an indication of whether the sub-criterion for that tunneling electrocardiogram is satisfied.

[0083] In some aspects, the techniques described herein relate to a method further including applying an appropriate implantation site machine learning model that inputs a location electrocardiogram and outputs an indication of whether the lead is at an appropriate implantation site.

[0084] In some aspects, the techniques described herein relate to a method wherein the appropriate implantation site machine learning model further inputs information relating to cardiac morphology.

[0085] In some aspects, the techniques described herein relate to a method further including applying an effective position machine learning model that inputs a tunneling electrocardiogram and outputs an indication of whether the lead is at an effective position.

[0086] In some aspects, the techniques described herein relate to a method wherein the implantation of the lead is for stimulating the His bundle, the left bundle branch, or the right bundle branch.

[0087] In some aspects, the techniques described herein relate to a method further including applying an assessment machine learning model that inputs a baseline electrocardiogram collected from the patient prior to an implantation procedure and outputs an assessment relating to presence of a physiological condition for which the implantation is a treatment.

[0088] In some aspects, the techniques described herein relate to a method further including applying a target zone machine learning model that inputs a cardiac image collected from the patient and outputs an indication of the target zone.

[0089] In some aspects, the techniques described herein relate to a method further including providing information to a catheter navigation system that control movement of the lead to the target zone.

[0090] In some aspects, the techniques described herein relate to one or more computing systems including: one or more computer-readable storage mediums that store computer-executable instructions for controlling the one or more computingsystems to: display a graphical representation of a heart a patient indicating a target zone for implantation of a lead of a pacemaker; identify locations of the lead based on electrocardiograms collected during pacing the lead as the lead is moved within the heart; display on the graphical representation an indication of the identified locations; determine, based on the morphology of the electrocardiogram at a location, whether it location is an appropriate implantation site; output an indication that the lead is at an appropriate implantation site; during tunneling of the lead within the heart wall, determine whether the lead is at an effective position within the heart wall based on tunneling electrocardiograms collected during tunneling; and output an indication that the lead is at an effective position; and one or more processors for controlling the one or more computing systems to execute one or more of the computer-executable instructions.

[0091] In some aspects, the techniques described herein relate to one or more computing systems wherein the graphical representation is based on a cardiac geometry that is derived from one or more images of the heart.

[0092] In some aspects, the techniques described herein relate to one or more computing systems wherein the computer-executable instructions include instructions that provide information to a catheter navigation system that control movement of the lead to the target zone.

[0093] In some aspects, the techniques described herein relate to one or more computing systems wherein the effective position is within the interventricular septum of the heart at a position at which pacing the lead stimulates the left bundle branch.

[0094] In some aspects, the techniques described herein relate to one or more computing systems including: one or more computer-readable storage mediums that store computer-executable instructions for controlling the one or more computing systems to: display a graphical representation of a heart of a patient indicating a target zone for implantation of a lead of a pacemaker; collect an electrocardiogram while pacing the lead at a location within the heart; determine, based on morphology of the electrocardiogram, whether the location is an appropriate site for implantation; output an indication that the lead is at an appropriate implantation site; during tunneling of the lead within the heart wall, collect a tunneling electrocardiogram and determine whether the lead is positioned at an effective position within the heart wall based on the tunnelingelectrocardiogram; and output an indication that the lead is at an effective position; and one or more processors for controlling the one or more computing systems to execute one or more of the computer-executable instructions.

[0095] In some aspects, the techniques described herein relate to one or more computing systems further including instructions that apply an appropriate implantation site machine learning model that inputs a location electrocardiogram and outputs an indication of whether the lead is at an appropriate implantation site.

[0096] In some aspects, the techniques described herein relate to one or more computing systems further including instructions that apply an effective position machine learning model that inputs a tunneling electrocardiogram and outputs an indication of whether the lead is at an effective position.

[0097] In some aspects, the techniques described herein relate to one or more computing systems further including instructions that apply an assessment machine learning model that inputs a baseline electrocardiogram collected from the patient prior to an implantation procedure and outputs an assessment relating to presence of a physiological condition for which the implantation is a treatment.

[0098] In some aspects, the techniques described herein relate to one or more computing systems further including instructions that apply a target zone machine learning model that inputs a cardiac image collected from the patient and outputs an indication of the target zone.

[0099] In some aspects, the techniques described herein relate to one or more computing systems further including instructions that provide information to a catheter navigation system that control movement of the lead to the target zone.

[0100] All documents incorporated by reference are incorporated in their entirety for the full extent of their disclosures. In the event of inconsistencies between the language in this document and any incorporated-by-reference document, the language in the incorporated-by-reference document should be considered supplementary to that of this document and the language in this document controls.

[0101] Although the subject matter has been described in language specific to structural features and / or acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts describedabove. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

Claims

CLAIMS1 . A method performed by one or more computing systems for guiding implantation of a pacemaker for physiological pacing of a heart of a patient, the method comprising: displaying a graphical representation of a heart that includes an indication of a target zone and an avoidance zone for implantation of the lead; for each of a plurality of location electrocardiograms collected while pacing the lead at locations within the heart, identifying a location of the lead based on the electrocardiogram; displaying on the graphical representation an indication of identified location; and determining whether the identified location corresponds to an appropriate implantation site based on the electrocardiogram; when the identified location is an appropriate implantation site, outputting an indication that the lead is at an appropriate implantation site; and for each of one or more of tunneling electrocardiograms collected during tunneling of the lead within the heart wall, determining whether the lead is at an effective position within the heart wall based on the tunneling electrocardiograms; and when the lead is at an effective position, outputting an indication that the lead is at an effective position.

2. The method of claim 1 wherein the determination of whether the identified location is an appropriate implantation site is based on an appropriate implantation site criterion that includes sub-criteria that is each based on morphology of a location electrocardiogram and further comprising, for each sub-criterion, displaying a voltagetime series representation of the location electrocardiogram for that sub-criterion with an indication of whether the sub-criterion is satisfied.

3. The method of claim 1 wherein the determining of whether the lead is at an effective position is based on an effective location criterion that includes sub-criteria that is each based on morphology of a tunneling electrocardiogram and further comprising,for each sub-criterion, displaying a voltage-time series representation of the tunneling electrocardiogram for that sub-criterion with an indication of whether the sub-criterion for that tunneling electrocardiogram is satisfied.

4. The method of claim 1 further comprising applying an appropriate implantation site machine learning model that inputs a location electrocardiogram and outputs an indication of whether the lead is at an appropriate implantation site.

5. The method of claim 4 wherein the appropriate implantation site machine learning model further inputs information relating to cardiac morphology.

6. The method of claim 1 further comprising applying an effective position machine learning model that inputs a tunneling electrocardiogram and outputs an indication of whether the lead is at an effective position.

7. The method of claim 1 wherein the implantation of the lead is for stimulating the His bundle, the left bundle branch, or the right bundle branch.

8. The method of claim 1 further comprising applying an assessment machine learning model that inputs a baseline electrocardiogram collected from the patient prior to an implantation procedure and outputs an assessment relating to presence of a physiological condition for which the implantation is a treatment.

9. The method of claim 1 further comprising applying a target zone machine learning model that inputs a cardiac image collected from the patient and outputs an indication of the target zone.

10. The method of claim 1 further comprising providing information to a catheter navigation system that control movement of the lead to the target zone.1 1 . One or more computing systems comprising: one or more computer-readable storage mediums that store computerexecutable instructions for controlling the one or more computing systems to: display a graphical representation of a heart a patient indicating a target zone for implantation of a lead of a pacemaker; identify locations of the lead based on electrocardiograms collected during pacing the lead as the lead is moved within the heart; display on the graphical representation an indication of the identified locations; determine, based on the morphology of the electrocardiogram at a location, whether it location is an appropriate implantation site; output an indication that the lead is at an appropriate implantation site; during tunneling of the lead within the heart wall, determine whether the lead is at an effective position within the heart wall based on tunneling electrocardiograms collected during tunneling; and output an indication that the lead is at an effective position; and one or more processors for controlling the one or more computing systems to execute one or more of the computer-executable instructions.

12. The one or more computing systems of claim 1 1 wherein the graphical representation is based on a cardiac geometry that is derived from one or more images of the heart.

13. The one or more computing systems of claim 11 wherein the computerexecutable instructions include instructions that provide information to a catheter navigation system that control movement of the lead to the target zone.

14. The one or more computing systems of claim 1 1 wherein the effective position is within the interventricular septum of the heart at a position at which pacing the lead stimulates the left bundle branch.

15. One or more computing systems comprising: one or more computer-readable storage mediums that store computerexecutable instructions for controlling the one or more computing systems to: display a graphical representation of a heart of a patient indicating a target zone for implantation of a lead of a pacemaker; collect an electrocardiogram while pacing the lead at a location within the heart; determine, based on morphology of the electrocardiogram, whether the location is an appropriate site for implantation; output an indication that the lead is at an appropriate implantation site; during tunneling of the lead within the heart wall, collect a tunneling electrocardiogram and determine whether the lead is positioned at an effective position within the heart wall based on the tunneling electrocardiogram; and output an indication that the lead is at an effective position; and one or more processors for controlling the one or more computing systems to execute one or more of the computer-executable instructions.

16. The one or more computing systems of claim 15 further comprising instructions that apply an appropriate implantation site machine learning model that inputs a location electrocardiogram and outputs an indication of whether the lead is at an appropriate implantation site.

17. The one or more computing systems of claim 15 further comprising instructions that apply an effective position machine learning model that inputs a tunneling electrocardiogram and outputs an indication of whether the lead is at an effective position.

18. The one or more computing systems of claim 15 further comprising instructions that apply an assessment machine learning model that inputs a baseline electrocardiogram collected from the patient prior to an implantation procedure andoutputs an assessment relating to presence of a physiological condition for which the implantation is a treatment.

19. The one or more computing systems of claim 15 further comprising instructions that apply a target zone machine learning model that inputs a cardiac image collected from the patient and outputs an indication of the target zone.

20. The one or more computing systems of claim 15 further comprising instructions that provide information to a catheter navigation system that control movement of the lead to the target zone.

Images