Robotically controlled intracardiac echocardiography (ICE) with artificial intelligence (AI) based analyses

Abstract

Systems and methods for a robotically controlled intracardiac echocardiography (ICE) with artificial intelligence (AI) based analyses. An example system includes a robotic device and an ICE catheter including a catheter body having a proximal end and a distal end and being adjustable in degrees of freedom of a tip portion of the ICE catheter. The robotic device controlling the degrees of freedom via drive mechanisms. The example system further includes processors that analyze ultrasound information being obtained via the ICE catheter, the ultrasound information reflecting a viewpoint associated with an interior of a heart, detect indicia of movement of the ICE catheter, and cause implementation of adjustments of individual degrees of freedom of the tip portion based on a patient mesh including vertices forming a contour associated with the heart.

Description

LAZA.046WO PATENTROBOTICALLY CONTROLLED INTRACARDIAC ECHOCARDIOGRAPHY (ICE) WITH ARTIFICIAL INTELLIGENCE (Al) BASED ANALYSESCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Prov. Patent App. No. 63 / 699620 filed September 26, 2024, and titled “ROBOTICALLY CONTROLLED INTRACARDIAC ECHOCARDIOGRAPHY (ICE) WITH ARTIFICIAL INTELLIGENCE (Al) BASED ANALYSES,” the disclosure of which is hereby incorporated herein by reference in its entirety.BACKGROUND

[0002] Intracardiac echocardiography (ICE) generally refers to ultrasound imaging performed from within cardiac chambers or vasculature using a catheter-borne probe positioned near the distal tip. ICE can acquire two-dimensional images, three- dimensional volumes, or time-resolved volumetric data and may provide near-field visualization of intracardiac structures, devices, and tissue responses during diagnostic or interventional workflows. A human operator, such as a medical professional, may control the ICE catheter while viewing a rendering of ultrasound information. With respect to control, the human operator may control degrees of freedom such as translation, rotation, flexion, and so on.SUMMARY

[0003] Example embodiments include systems, methods, and computer readable media. An example system includes a robotically controlled intracardiac echocardiograph (ICE) system. The robotically controlled ICE system includes a robotic device comprising a housing comprising a recess having one or more drive mechanisms disposed therein; an ICE catheter comprising a catheter body having a proximal end and a distal end, an imaging probe disposed at a tip portion of the catheter body disposed adjacent to the distal end, and a control handle disposed at the proximal end, the controlhandle having one or more actuators configured to adjust one or more degrees of freedom of the tip portion of the ICE catheter; wherein the recess is configured to receive at least a portion of the control handle such that the one or more drive mechanisms engage the one or more actuators, wherein when engaged, the one or more drive mechanisms can actuate the one or more actuators to adjust one or more degrees of freedom of the tip portion of the ICE catheter; and one or more processors configured to execute instructions that cause the one or more processors to: analyze ultrasound information being obtained via the ICE catheter, the ultrasound information reflecting a viewpoint associated with an interior of a heart; detect indicia of movement of the ICE catheter; and cause implementation of one or more adjustments of individual degrees of freedom of the tip portion based on a patient mesh including a plurality of vertices forming a contour associated with the heart, wherein the adjustments are configured to update a pose associated with the ICE catheter to correspond with the viewpoint.

[0004] Another example system includes a robotically controlled intracardiac echocardiograph (ICE) system. The robotically controlled ICE system includes a robotic device comprising one or more drive mechanisms. The robotically controlled ICE system includes an ICE catheter comprising a catheter body having a proximal end and a distal end, an imaging probe disposed at a tip portion of the catheter body disposed adjacent to the distal end. The ICE catheter includes a control handle disposed at the proximal end, the control handle has one or more actuators configured to adjust one or more degrees of freedom of the tip portion orthe imaging probe of the ICE catheter. The robotic device is configured to engage the one or more drive mechanisms with the one or more actuators. When engaged, the one or more drive mechanisms can actuate the one or more actuators to adjust one or more degrees of freedom of the tip portion orthe imaging probe of the ICE catheter. The robotically controlled ICE system also includes one or more processors configured to execute instructions that cause the one or more processors to: analyze ultrasound information being obtained via the ICE catheter, the ultrasound information reflecting a viewpoint associated with a a heart; detect indicia of movement of the ICE catheter; and cause implementation of one or more adjustments of individual degrees of freedom of the tip portion orthe imaging probe based on a modelof a contour associated with the heart, wherein the adjustments are configured to update a pose associated with the ICE catheter to correspond with the viewpoint.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 illustrates an example system environment for a robotic device used in an imaging procedure.

[0006] Figure 2 illustrates a block diagram of generating a patient mesh based on an input of, at least, intracardiac echocardiography (ICE) ultrasound information.

[0007] Figure 3 is a flowchart of an example stabilization process for a robotically controlled ICE catheter.

[0008] Embodiments of the present disclosure and their advantages are best understood by referring to the detailed description that follows. It should be appreciated that the figures are for purposes of illustrating embodiments of the present disclosure and not for purposes of limiting the same.DETAILED DESCRIPTIONIntroduction

[0009] The disclosed technology relates to a robotically controlled intracardiac echocardiography (ICE) catheter. ICE, as described herein, may encompass different types and modes of imaging, such as 2D, 3D, 4D, ultrasound imaging. As will be described, the ICE catheter may be autonomously or semi-autonomous controlled to navigate to disparate cardiac views. The ICE catheter may leverage artificial intelligence techniques to enable such autonomous or semi-autonomous control, including techniques to address disturbances or motion which have negatively affected use of ICE priorto the disclosed technology. In some embodiments, a same system or housing may robotically control the ICE catheter in combination with, or separately from, a transesophageal echocardiogram (TEE) probe. Example description related to a robotically controlled TEE probe and / or robotically controlled ICE catheter is included in International Application No. PCT / US2023 / 061573, which is hereby incorporated hereinby reference in its entirety.

[0010] As an example, ICE refers to ultrasound imaging performed from within the heart using a catheter-borne probe that acquires images from a vantage point near the distal tip of the catheter. As known by those skilled in the art, ICE may be used in structural heart and electrophysiology procedures, such as left atrial appendage occlusion, tricuspid Transcatheter Edge-to-Edge Repair (TEER), mitral valve repair and / or replacement procedures and so on. Advantageously ICE may provide enhanced imaging as compared to other technologies (e.g. TEE) such as reduced shadowing. Because the ICE catheter operates intravascularly and intra-chamber, ICE can supply substantially real-time views of anatomy at close range. Additionally, ICE can supply substantially real-time views of devices utilize in cardiac procedures (e.g., interventional devices or other devices). This can, as an example, enhance decision making during such cardiac procedures while enhancing patient safety via monitoring distances between such devices and a cardiac surface of the patient.

[0011] Due to the size of the ICE catheter, and its location within a patient’s body, precise control of the ICE catheter is paramount for safety. Automated robotic control provides such precise control and positioning of the catheter, for example via fine control of degrees of freedom of the ICE catheter. Example degrees of freedom may include flexion, rotation, translation, and so on. Additionally, robotic control enables substantially improved stabilization of the ICE catheter which guards against accidental disturbances that may bump or adjust the degrees of freedom of the catheter. As known by those skilled in the art, at present medical professionals may leverage stabilization techniques such as a wet towel to help limit movement of the ICE catheter which controls the degrees of freedom of the distal tip (e.g., the wet towel may help guard against the catheter from rolling on a bed of a patient).

[0012] The robotic control described herein, such as implemented by a robotic device (e.g., robotic device 102), may leverage example mechanical and software elements to ensure positioning and stabilization. For example, the robotic control may receive a handle of an ICE catheter (e.g., via a holster or other element comprising a housing or a holder configured to receive the handle). The handle may include one or more actuators, such as dials, knobs, or control levers. In this example, the one or moredials, knobs, or control levers or elements, may be manipulated via the robotic device. As one example, the robotic device may receive the handle and may include motorized control elements that couple with the dials. The robotic device may activate the motorized control elements to similarly precisely control adjustments to the dials or other actuators and monitor for any slipping, deviations of movement, and so on. Example description related to such robotic control of dials is included in International Patent Publication WO 2025 / 090965 and WO 2025 / 144824 which is incorporated herein by reference in their entirety. While these patent publications relate to TEE probes, similar techniques may be applied to ICE catheters.

[0013] As described herein, the robotic control may address disturbances caused by patient movement (e.g., inadvertent movement, and so on), operator or medical professional movement, and so on. For example, a system (e.g., console system 108) may analyze visual information being obtained from the distal tip of the ICE catheter. In this example, the visual information may reflect 2D, 3D, 4D, ultrasound information. Based on rapid adjustments in view, for example using optical flow or other computer vision or machine learning based techniques, the system may ascertain that the ICE catheter was bumped or adjusted. The system may leverage information informing movement due to cardiac phase (e.g., an ECG) or breathing and determine that the adjustment in view was not due to natural patient movement.

[0014] To address such disturbances, the system may determine adjustments, for example micro-adjustments, to return the ICE catheter to a pose associated with a prior viewpoint (e.g., prior to being bumped or adjusted). In some embodiments, pose may reflect orientation and translation. Pose may be indicative of the degrees of freedom. Advantageously, the system may monitor substantially continuously for adjustments. As one example, the system may periodically, or substantially continuously, effectuate micro-adjustments of the ICE catheter dials to ensure a stable viewpoint is maintained. The stable viewpoint may correspond, for example, to a particular cardiac view of interest to a medical professional or operator. The stable viewpoint may also correspond, for example, to a particular cardiac view that forms part of a workflow being performed.

[0015] With respect to different cardiac views, the ICE catheter may be controlled to navigate between these cardiac views. For example, an operator or medical professionalmay select specific cardiac views via a user interface. The operator or medical professional may additionally be implementing a workflow associated with a procedure. For this example, the ICE catheter may be navigated to cardiac views that form at least part of the workflow.

[0016] In some embodiments, a heart model (referred to herein as a graph model or patient mesh) may be generated, at least in part, based on ultrasound information from the ICE catheter. This heart model may be leveraged, in some embodiments, to inform navigation between these cardiac views. In some embodiments, the heart model may reflect an adjusted or deformed average heart model associated with a multitude of patients. For example, the average heart model may include a mesh, or other geometry information, that includes vertices which form the contour of a heart. The ICE catheter may obtain ultrasound information, for example 3D or 4D information, and the system described herein may generate the heart model based on the view of the patient’s heart and the average heart model. As may be appreciated, the ICE catheter may obtain ultrasound information in which the apex is positioned at the distal tip of the catheter. As the ICE catheter maneuvers about, the ICE catheter can view additional portions of the interior of the heart. Thus, the heart model may be refined during this movement to provide for an accurate representation of the patient’s heart.

[0017] In some embodiments, the heart model may initially be generated, at least in part, using TEE ultrasound information which may be able to view a larger portion of the heart. In some embodiments, the heart model may be initially generated, at least in part, using a computed tomography (CT) scan. The system may then adjust this heart model based on the interior view being obtained via the ICE catheter.

[0018] Certain example technological challenges with ICE may be addressed using the techniques described herein. For example, typically two or more operators may be required for ICE. In this example, a first operator may control movement of an interventional catheter or device while a second operator may control movement of the ICE catheter, probe or imaging aspects thereof. Using view-based controls, such as the user interface described above, and other approaches described herein, a single operator may be equipped to manipulate the interventional catheter or device and the ICE catheter, probe or imaging aspects thereof. An example of view-based control isdescribed in, for example, International Publication No. WO 2023 / 1 7544.

[0019] As one example described herein, robotically controlled ICE may cause movement of the catheter to obtain specific views of interest to the operator. Asa further example, robotically controlled ICE may employ a catheter sized and shaped for percutaneous insertion into the patient. The catheter has an imaging probe with a field of view, coupled to the catheter near a distal end thereof. Robotically controlled ICE also may employ a drive mechanism (e.g., mechanism 120) and a processor (e.g., processor 110, 130, 140). The drive mechanism can be coupled to the catheter and / or the imaging probe and can be configured to control, e.g., translate and / or rotate, the imaging probe. The processor can be operatively coupled to the imaging probe and the drive mechanism and can exchange (e.g., transmit and receive) signals with the drive mechanism and the imaging probe. The processor can control the drive mechanism to place the imaging probe at a first position and at a second position within the patient, as discussed further below. The processor can control the imaging probe to generate first image data and second image data related to respective first and second fields of view therefrom. The processor can be configured to generate and / or update a model of the anatomical structure, e.g., as discussed further below, for example considering the first and second image data. Robotically controlled ICE may additionally reduce navigation time, for example to more rapidly perform septal crossing.

[0020] As known by those skilled in the art, ICE catheter instability presents large technological hurdles. For example, the ICE catheter may get bumped, nudged, bent, and so on, which causes rapid adjustment of the view seen by an operator. This may also result in higher complication results due to objects of interest being out of view and / or the tip of the ICE catheter uncontrollably impacting the anatomy. In contrast, the robotically controlled ICE may enhance stability through use of a holster or other element to maintain catheter control. If the catheter is knocked, or otherwise displaced or adjusted, the robotically controlled ICE may use artificial intelligence (Al) techniques to return to a desired view.

[0021] With respect to the patient model, the system may create the model based on ultrasound information and, in some embodiments, the above-described average model. The average model of the heart may be based on the size and or shape of relevantstructures of the hearts of a plurality of individual patients. In this example, once the average model of the heart has been established, the average model may be deformed or otherwise refined based on the ultrasound images of a particular patient being treated to more accurately provide a model the heart of the particular patient, or a patient specific model. In some embodiments, the system may perform deformation via a forward pass through a graph convolutional network (GCN) or other machine learning model. Based on a current view being obtained from the ultrasound, the system may determine where the ICE catheter is positioned based on the model. Thus, the system may determine how to navigate to and / or return to the desired view. The patient specific model may be incrementally built by deforming or otherwise refining the average model based upon each image captured by the imaging probe. Some aspects of the patient specific heart model may be estimated, e.g., where the shape or size is not needed to perform a step or procedure and / or where no image data has yet been obtained. For example, the average heart model may be defined by a 3D array of nodes. A 3D surface of a patient specific heart model may be estimated from one or more images or measurements obtained at positions away from the area of the surface being estimated. For example, the location of the apex of the left ventricle may be indicative of a surface extending therefrom and may provide a basis for estimating a surface or portion of the heart even the imaging probe has not captured an image of that surface.

[0022] Artificial intelligence (Al), or optionally classical techniques, may additionally be used for image assessments. For example, automated LAA measurements may be automatically determined. Useful information may be automatically presented to an operator, for example via a user interface. In some embodiments, the system may determine whetherthe ICE catheter is folding or bending such that it is not substantially straightened out. For example, the system may determine (e.g., based on the abovedescribed model) that views being obtained via the ultrasound element are not corresponding to expected views based on movement of the catheter. As an example, the operator may indicate that the catheter is to progress forward. For this example, the system may determine that the views being obtained are not indicative of forward movement. Thus, the system may determine that the catheter is bending and may cause the catheter to be pulled back or otherwise adjusted in orientation.

[0023] In some embodiments, the system may correlate instructed robotic adjustments of the ICE catheter with the resulting visual data from the imaging probe (e.g., the ultrasound transducer at the tip). For instance, if the system instructs the drive mechanism to advance the catheter by 5 mm, an Al image analysis technique may expect a corresponding forward shift in the visual field, characterized by a predictable optical flow. If the technique detects minimal or incongruous changes in the visual field despite the commanded advancement, the system infers that the catheter is encountering resistance and bending. Upon this determination, it can automatically halt forward movement and alert the operator.

[0024] The above and other aspects will now be described in more detail.

[0025] Figure 1 illustrates an example system environment 100 for a robotic device used in an imaging procedure. In one or more embodiments, the system environment 100 may be used to perform imaging of a patient anatomy during a medical procedure. The system environment 100 may include one or more of a robotic device 102, an ICE catheter 104, one or more other devices 106, and a console system 108. While this disclosure presents a system environment 100 with a number of different components arranged in a specific configuration, this disclosure contemplates a system environment 100 with any suitable number of different components arranged in any suitable configuration. As an example and not by way of limitation, the system environment 100 may include an additional imaging system, such as a TEE probe that is connected to the robotic device 102 or to a separate robotic device not shown. As another example and not byway of limitation, the system environment 100 may not include one or more other devices 106. In one or more embodiments, the robotic device 102 may be physically coupled to the ICE catheter 104 to manipulate the ICE catheter 104. As an example and not by way of limitation, the robotic device 102 may be coupled to one or more controls (e.g., knobs, dials) through one or more drive mechanisms (e.g., a motor) to maneuver the ICE catheter 104. In one or more embodiments, the robotic device 102 may be communicatively coupled to one or more of the one or more other devices 106 or a console system 108.

[0026] In one or more embodiments, the robotic device 102 may be used to maintain a position of the ICE catheter 104 relative to a patient anatomy. In one or moreembodiments, the robotic device 102 may be used to maintain a view of the ICE catheter 104. In one or more embodiments, the robotic device 102 may use one or more stabilization techniques to maintain a view presented by the ICE catheter 104. In one or more embodiments, the robotic device 102 may include one or more of a processor 110, storage 112, memory 114, communication unit 116, sensors 118, drive mechanism 120, and a data analysis module 122. While this disclosure describes the robotic device 102 as having a certain number of components in a particular arrangement, this disclosure contemplates the robotic device 102 as having any number of components in any suitable arrangement. In one or more embodiments, the robotic device 102 may store one or more machine-learning models (not shown), one or more models of patient anatomy, one or more models of devices, a plurality of data collected from one or more sensors 118, and other data sources in the storage 112.

[0027] The robotic device 102 may use processor 110 to execute one or more instructions in memory 114 to perform one or more functions of the robotic device 102 as described herein. In one or more embodiments, the robotic device 102 may communicate with the other components of the system environment 100 through communication unit 116. In one or more embodiments, the one or more sensors 118 may comprise one or more of an accelerometer, a gyroscope, a magnetometer, an electromagnetic tracker, and other sensors of a robotic device. In one or more embodiments, the sensors 118 may be embodied as an integrated sensor array. The sensors 118 can be configured as a position sensor, e.g., by processing data from one or more of the foregoing sensor types. The sensors 118 can be configured as a load sensor, e.g., by processing data from one or more of the foregoing sensor types.

[0028] In one or more embodiments, the robotic device 102 may use a drive mechanism 120 to manipulate the ICE catheter 104. As an example and not by way of limitation, the robotic device 102 may use the drive mechanism 120 to turn a knob of the ICE catheter 104 to adjust a position of the ICE catheter 104. In one or more embodiments, the data analysis module 122 may analyze the data received from the robotic device 102. As an example and not by way of limitation, the robotic device 102 may receive image data from the ICE catheter 104 and input the image data into the data analysis module 122 to determine one or more adjustments (e.g., micro-adjustments)that is needed by the ICE catheter 104 to maintain a preselected view. The data analysis module 122 may output one or more instructions for the drive mechanism 120 to manipulate the ICE catheter 104. In one or more embodiments, the data analysis module 122 may access data in storage 112 to perform an analysis using a machine-learning model and / or computer vision algorithms to generate one or more instructions for the drive mechanism 120.

[0029] While this disclosure describes the robotic device 102 as performing one or more functions to stabilize the ICE catheter 104 to present a particular view, this disclosure contemplates the robotic device 102 performing any suitable functions to stabilize the ICE catheter 104. As an example and not by way of limitation, the console system 108 may perform the data analysis and generation of instructions for the drive mechanism 120 to send instructions to the robotic device 102.

[0030] In one or more embodiments, the ICE catheter 104 may be used to perform an imaging process of patient anatomy using ultrasound imaging. In one or more embodiments, the ICE catheter 104 may include one or more transducers 124, sensors 126, and an input / output (I / O) interface. In one or more embodiments, the ICE catheter 104 may use one or more transducers 124 to gather imaging data. In one or more embodiments, the ICE catheter 104 may use one or more sensors 126 to measure one or more of a pressure, an acceleration, and the like experienced by the ICE catheter 104. In one or more embodiments, the sensors 126 may include one or more of a pressure sensor, an accelerometer, a gyroscope, a magnetometer, and other sensors of an ICE catheter 104. In one or more embodiments, the sensors 126 may be embodied as an integrated sensor array. The sensors 126 can be configured as a position sensor, e.g., by processing data from one or more of the foregoing sensor types. The sensors 126 can be configured as a load sensor, e.g., by processing data from one or more of the foregoing sensor types. In one or more embodiments, the ICE catheter 104 may use an I / O interface 128 to transmit data to one or more of the robotic device 102 or console system 108. In one or more embodiments, the ICE catheter 104 may receive one or more physical manipulations from the robotic device 102 to control a position, pose, or configuration of the ICE catheter 104. For example, the ICE catheter 104 can include a control handle disposed at a proximal end of a catheter body thereof. The control handle can have oneor more actuators 129 configured to adjust one or more degrees of freedom of a tip portion of the catheter body of the ICE catheter. The tip portion can be disposed at a distal end of the catheter body. The one or more actuators 129 can manipulate the tip portion to provide flexion of the tip portion relative to a more proximal portion of the catheter body. The one or more actuators 129 can manipulate the transducer to rotate the transducer about the longitudinal axis of the catheter body. The one or more actuators 129 can manipulate the catheter body with the tip portion straight or flexed. In one or more embodiments, the one or more physical manipulations may be used to maintain a preselected view of the ICE catheter 104. While this disclosure describes the ICE catheter 104 as performing one or more functions for imaging patient anatomy, this disclosure contemplates the ICE catheter 104 performing any suitable function for imaging patient anatomy.

[0031] In one or more embodiments, the other devices 106 may be used for a medical procedure. In one or more embodiments, the other devices 106 may include one or more of a ventilator, an anesthesia machine, EKG monitor, blood pressure monitor, and other devices used in a medical procedure. In one or more embodiments, the other devices 106 may use one or more sensors 136 to measure data during the medical procedure and store the measured data in storage 132. In one or more embodiments, the processor 130 may execute instructions stored in the memory 134 to perform one or more functions of the other devices (e.g., providing oxygen to the patient, measure blood pressure, etc.). In one or more embodiments, the I / O interface 138 may be used to transmit data collected by the other devices 106 to one or more of a robotic device 102 or a console system 108. Other devices 106 may affect the physiology of the patient, such as a ventilator providing oxygen to the user. In one or more embodiments, other devices 106 may provide data that may be used by the data analysis module 122, 148 to generate one or more instructions for the drive mechanism 120 to compensate for the effect the other devices 106 has on patient physiology. As an example and not by way of limitation, a ventilator may send data of how much oxygen it's providing to a patient that is used by the data analysis module 148 to determine a projected movement by the patient in response to the provided oxygen. The data analysis module 148 may use a machine-learning model and / or computer vision algorithms to determine an estimated movement of an ICEcatheter 104 associated with a plurality of data sources and subsequent instructions to send to the drive mechanism 120 to compensate for the estimated movement to maintain a preselected view of the ICE catheter 104.

[0032] In one or more embodiments, the console system 108 may present a current view of patient anatomy received by the ICE catheter 104. In one or more embodiments, the console system 108 may perform data analysis using one or more data sources to determine instructions for the drive mechanism 120 of the robotic device 102 to adjust the ICE catheter 104 to maintain a preselected view. In one or more embodiments, console system 108 may include one or more of a processor 140, a memory 142, a storage 144, a communication unit 146, a data analysis module 148, a user interface 150, and a display 152. In one or more embodiments, the processor 140 may execute instructions stored in memory 142 to perform one or more functions of the console system 108. In one or more embodiments, the console system 108 may store data collected from one or more sources, one or more models of patient anatomy, and one or more models of devices (e.g., devices to be implanted in patient) in storage 144. In one or more embodiments, the console system 108 may use data analysis module 148 similarly to data analysis module 122 to process data to generate one or more instructions to the drive mechanism 120 of the robotic device 102 to control the ICE catheter 104.

[0033] In one or more embodiments, a user interface 150 may receive user inputfrom one or more operators. As an example and not byway of limitation, the user interface 150 may receive a user input of a selected view an operator would like to see from the ICE catheter 104. In one or more embodiments, the selection of the user input may generate instructions to send to the drive mechanism of the robotic device 102 to manipulate the ICE catheter 104 to location with respect to the patient anatomy to present the selected view to the operator. In one or more embodiments, display 152 may display a current view selected by the operator of the console system 108.

[0034] In one or more embodiments, the console system 108 may use one or more models of the patient anatomy and one or more models of devices with the machinelearning model and / or computer vision algorithms to generate instructions for the drive mechanism 120 to manipulate the ICE catheter 104 to maintain a preselected view. Inone or more embodiments, the user interface 150 may receive a user input indicative of a workflow selected by the operator, where the workflow specifies an order of one or more views to be presented by the ICE catheter 104. In one or more embodiments, the console system 108 may generate instructions for the drive mechanism 120 of the robotic device 102 to manipulate the ICE catheter 104 to position the ICE catheter 104 in a sequence of locations with respect to the patient anatomy to obtain a sequence of views associated with the workflow. While this disclosure describes a console system 108 as performing one or more functions for presenting patient anatomy, this disclosure contemplates a console system 108 performing any suitable function for presenting patient anatomy.

[0035] As described herein, the robotic device 102 may use stabilization techniques to maintain a pose of the ICE catheter 104. For example, an operator may select a particular view to be displayed based on the ultrasound information being obtained via the ICE catheter 104, where the robotic device 102 may position the ICE catheter in an initial orientation based on a predetermined orientation established with the selected view. As ay be appreciated, maintaining the image presented to the operator is critical because the operator may be performing an intensive procedure on a patient that requires precise imaging to properly navigate the patient anatomy. As such, while external events may occur to disturb the imaging process, the robotic device 102 may implement stabilization techniques so that the ICE catheter 104 can maintain the selected view.

[0036] In some embodiments, the robotic device 102 may perform microadjustments to one or more control elements of the ICE catheter to maintain a fixed pose in relation to the patient anatomy or in relation to another device by centering the other device in view. Example control elements may include dials, knobs, and so on. The robotic device 102 may additionally control translation of the ICE catheter such that micro-adjustments may include adjustments in the ICE catheter’s position within the patient.

[0037] As an example and not by way of limitation, the console system 108 may be displaying a preselected view of a mitral valve of the heart of the patient at a particular angle (e.g., based on the ICE catheter 104), and an external event (e.g., a disturbance,nudge, bump, etc.) may be introduced to the system. As another example and not byway of limitation, the console system 108 may be displaying a mitral valve clip centered in the presented view (e.g., via the display 152). Continuing the preselected view of a mitral valve example, to maintain the original preselected view of the mitral valve, the robotic device 102 may adjust a joint of the ICE catheter to accommodate the movement that is experienced from the external event. To do so, the robotic device 102 and / or console system 108 may receive ultrasound information from the ICE catheter and analyze, using machine-learning models and / or computer vision techniques, the image data to determine a motion that is experienced by the imaging probe in relation to the patient anatomy. The determined motion may then be used as an input into machine-learning models and / or computer vision techniques to generate one or more micro-adjustments needed to be performed by the robotic device 102 to move the ICE catheter 104 to the initial pose corresponding to the preselected view. In one or more embodiments, a fixed location of the ICE catheter 104 may change as an operator proceeds through a workflow. As an example and not by way of limitation, if an operator is proceeding through a workflow to implant a mitral valve clip, then the ICE catheter may transition from a fixed position in relation to the septum, to a fixed position in relation to the mitral valve, and subsequently to another fixed position in relation to an implanted mitral valve clip.

[0038] In one or more embodiments, the robotic device 102, or console system 108, may track one or more devices and relative distances between models of the one or more devices and a model of the patient anatomy (e.g., the heart model described herein). The robotic device 102, or console system 108, may use machine-learning models and / or computer vision techniques to perform the device tracking and track the distances between different models (e.g., device model to heart model). The robotic device 102, or console system 108, may use known information, such as a model of a device (e.g., a model of a mitral valve clip) and the heart model to identify, or otherwise determine, when tissue interaction occurs (e.g., a device is touching the heart of the patient). The robotic device 102, or console system 108, may perform this tracking despite one or more of the device or the patient anatomy being occluded or outside the field of view of the ICE catheter. As an example and not by way of limitation, the robotic device may trackthe location of a device within a patient anatomy using a model of the device (e.g., a mitral valve clip) and a model of the patient anatomy (e.g., the heart model of the patient).

[0039] In one or more embodiments, additional data may be used by the machinelearning models and / or computer vision techniques to perform one or more of a determination of the motion that is experienced by the imaging probe, a generation of one or more micro-adjustments needed to be performed by the robotic device 102 to move the ICE catheter, and other functions of the robotic device 102. The additional data may include data inputs from one or more other devices (e.g., a ventilator, electrocardiogram (ECG), and so on), one or more sensors coupled to the robotic device 102, and other sources of additional data for stabilization techniques used by the robotic device 102. In one or more embodiments, one or more sensors may be coupled to one or more parts of the robotic device 102 and / orthe ICE catheter 104. As an example and not by way of limitation, a pressure sensor may be integrated into the imaging probe of the ICE catheter 104, an accelerometer or gyroscope may be integrated into the imaging probe of the ICE catheter 104, and the like. The robotic device 102 and / or console system 108 may receive the sensor data, such as accelerometer data and use the accelerometer data as an input into machine-learning models and / or computer vision algorithms to determine a motion of the ICE catheter. While this disclosure describes a particular embodiment of using one or more stabilization techniques to maintain a position of an ICE catheter, this disclosure contemplates other embodiments of using one or more stabilization techniques to maintain a position of an ICE catheter.

[0040] Figure 2 illustrates a block diagram of generating a patient mesh 220 based on an input of, at least, intracardiac echocardiography (ICE) ultrasound information 200. The elements of Figure 2 may be implemented via the system environment 100 described above. In some embodiments, a system (e.g., the console system 108, or optionally robotic device 102) may generate the patient mesh 220. As described herein, the patient mesh 220 may represent an example of a heart model of a patient. For example, the patient mesh 220 may include a multitude of vertices in three-dimensional space which define, or inform, the contour or shape of the patient’s heart.

[0041] In the illustrated embodiment, a data augmentation engine 202 obtains ultrasound information 200 via an ultrasound sensor positioned at a distal end of an ICEcatheter. In some embodiments, the ICE catheter may obtain 2D ultrasound images, 3D ultrasound information which may be time-resolved to form 4D ultrasound information. The system may access training data that includes ground truth labeled ultrasounds information. As may be appreciated, this engine 202 may be used during training (e.g., only used during training), for example to expand upon the training distribution while preserving label consistency.

[0042] In the illustrated embodiment, a 3D ultrasound volume 204 is depicted as extending from a probe. This volume 204 may represent an example of training data, and may, as an example, represent voxels. The data augmentation engine 202 may supplement this example through augmentation which herein is referred to as vascular augmentation 206. This augmentation 206 applies catheter-centric adjustments that emulate partial field-of-view conditions typical of intracardiac imaging while maintaining an anatomically plausible imaging origin. For example, vascular augmentation 206 may retain the near-field region at the imaging apex and vary the distal extent and sector aperture to simulate depth-limited views, optionally with small rotations about the tip frame and subtle intensity scaling to reflect routine manipulations and gain changes. In the illustrated example, a plane is illustrated as extendin through the distal extent. This plane may reflect an exterior, surface, or near exterior, of the patient’s heart. Additional effects such as controlled occlusions or near-field attenuation may be included to expose the learning stack to realistic visibility loss while remaining label-preserving, consistentwith the role of augmentation to increase generalizability from limited groundtruth datasets. Engine 202 may also layer probe-centric and affine transformations so that both pose and appearance diversity are represented in the augmented volumes.

[0043] A feature extraction engine 208 may be used to output hierarchical features 210. In some embodiments, the feature extraction engine 208 may represent a machine learning model, such as a U-NET, Swin-UNET, and so on. As known by those skilled in the art, U-NET may output a feature pyramid which includes a threshold number of feature maps at different scales. These hierarchical features 210, as will be described, may be used to deform the average heart mesh 216 (e.g., average heart model) described herein to form the patient mesh 220.

[0044] With respect to U-NET, hierarchical features 210 may be obtained from the encoder or decoder portion of U-NET. In some embodiments, the features 210 may represent features output via the encoder. However, in some embodiments the features 210 may represent features output via the decoder and / or a concatenation of the encoder and decoder features.

[0045] As illustrated, the engine 208 may determine a confidence map 212 associated with the hierarchical features 210. For example, features from the hierarchical features 210 may be sampled at individual voxel positions. The engine 208 may include a head that outputs individual confidence values or measures for features sampled at individual voxel positions. These confidence values or measures may indicate a confidence that the features accurately characterize a 3D ultrasound. For example, the features 210 may be associated with per-voxel confidence values learned to reflect the reliability of ultrasound-derived information. In some embodiments, during later stages, higher-confidence regions may carry greater influence and low-confidence regions may be attenuated, improving robustness to artifacts and partial views common to ICE acquisition.

[0046] The above-described confidence map 212 may be used, in some embodiments, to perform pre-alignment of the average heart mesh 216. For example, an alignment engine 214 may use the confidence map 212 optionally along with a correspondence field reflecting, for each vertex of the average heart mesh 216, a predicted 3D position within an input 3D ultrasound. The predicted 3D positions forms a correspondence field that maps each mesh vertex to a target anatomical location in the 3D ultrasound.

[0047] During pre-alignment, the engine 214 may determine a transformation between the average heart mesh 216 and the predicted 3D positions. Example transformations may include a rigid transformation, similarity transformation, affine transformation, and so on. These transformations may be weighted according to the confidence map 212, such that more trustworthy or confident predicted 3D positions are given greater influence in aligning corresponding vertices of the average heart mesh 216.

[0048] In some embodiments, for ICE-specific operation, engine 214can incorporate an intravascular tip prior (e.g., pose prior) that constrains candidate solutions. Forexample, the solutions may be constrained such that the imaging apex remains within a vascular envelope, optionally augmented by catheter pose cues, thereby improving stability and plausibility of the pre-alignment in the presence of limited field-of-view.

[0049] In some embodiments, the average heart mesh 216 may provide a labeled geometric prior that confers anatomical structure and persistent vertex identities to the downstream model. As one example, the mesh 216 may be derived from centraltendency statistics over a set of patient meshes, with each vertex carrying a region label such that, after deformation, patient mesh 220 may inherently be segmented at the vertex level (e.g., optionally without separate labeling steps). This labeling enables consistent correspondence across patients and supports feature projection, alignment, and loss computation tied to anatomically meaningful regions.

[0050] A graph model engine 218 may then output a patient mesh 220 based on the hierarchical features 210 and aligned average heart mesh 216. In some embodiments, the features 210 may be projected onto the average heart mesh 216. For example, interpolation may be used (e.g., trilinear interpolation). In this example, based on spatial locations of each vertex of the average heart mesh 216, the engine 218 may sample the hierarchical features 210 at the corresponding spatial locations. In this way, each vertex may be assigned features (e.g., a feature vector). In some embodiments, the engine 218 may use an attention-based technique to project features. For example, features for each vertex may be generated using attention with respect to the features 210.

[0051] The graph model engine 218 may represent a machine learning model, for example a graph convolutional network. In some embodiments, Chebyshev convolutions may be used. In some embodiments, the graph convolutional network may reflect a DiffusionNet. To output the patient mesh 220, the engine 218 may include a threshold number of graph blocks that iteratively cause deformation of the average heart mesh 216.

[0052] For example, a first block may deform the average heart mesh into a first intermediate mesh. In this example, a second block may deform the first intermediate mesh into a second intermediate mesh. There may be 2, 3, 4, 5, and so on, iterative blocks. In some embodiments, each block may leverage particular hierarchical features. For example, lower resolution features may be used in the first block while higherresolution features may be used in the second block. At each block, the features may be projected onto the mesh so that the mesh may be further deformed towards the patient mesh.

[0053] For ICE-oriented inference, the graph model engine 218 may leverage the confidence map 212 to gate per-vertex updates and reweight regularization in regions that are frequently out of view. This may help prevent overfitting to uncertain observations while allowing more confident anatomy to drive local corrections. In some embodiments, the engine 218 may apply occlusion-aware masks derived from the ICE sector geometry so vertices lacking line-of-sight evidence are influenced primarily by smoothness and anatomical priors, whereas vertices intersecting high-confidence features may receive greater update magnitudes.

[0054] Additionally, the engine 218 may initialize from a previously estimated patient mesh 220 when streaming frames are available, providing temporal consistency during routine catheter motion and rapid stabilization after disturbances. Thus, the engine 218 may iterate on the patient mesh 220 as the ICE catheter maneuvers about the patient’s heart.

[0055] In some embodiments, the engine 218 may identify key anatomical landmarks from the ultrasound image data, such as the mitral valve annulus, the fossa ovalis, left atrial appendage ostia, tricuspid valve annulus, ventricular outflow tracts, or the pulmonary vein ostia. By matching these identified landmarks to corresponding points on the average heart model, the engine 218 may compute a patient-specific transformation matrix. This matrix may then be applied to the average model to deform it, creating the patient mesh 220.

[0056] In some embodiments, the heart model may be used to inform recognition of structures (e.g., cardiac structures or locations). The heart model may additionally be used for localization of the ICE catheter, for example to provide information regarding questions such as what is in the field of view, where is the tip of the catheter (e.g., the ultrasound transducer) relative to the anatomy, and so on. An example processing pipeline that leverages a heart model follows.

[0057] The system, such as the console system 108, may receive substantially realtime ultrasound information (e.g., ultrasound video, images, 2D frames, 2.5D stacks, or 3D volumes, and so on) from the ICE catheter. The system may compute a forward pass through a deep learning model, such as a convolutional neural network (U-Net), attention-based network, and so on. The deep learning model may be trained based on labeled cardiac ultrasound images or information, such as from ICE catheters. In some embodiments, the deep learning model may be trained to emit per-pixel labels optionally along with confidence information. The system may thus generate semantic cues needed to distinguish walls, valves, device edges, and other landmarks while optionally providing uncertainty that downstream modules can use to temper decisions in artifact- prone regions.

[0058] The system may additionally perform tracking. For example, the system may group labeled elements (e.g., labeled pixels) into anatomical elements (e.g., mitral valve, atrial wall, LAA, and so on) and maintain their identities across time. In some embodiments, the system may fuse optical flow, learned correspondence fields, and / or motion models tied to cardiac and respiratory cycles, enabling robust persistence through partial occlusions or brief view loss. Instance states can include centroid, contour, velocity, and phase-aligned motion predictions, which together stabilize labels and produce temporally coherent trajectories for recognizable structures.

[0059] With respect to tracking, the system may therefore, as an example, not just see a cloud of mitral valve pixels but recognize the entirety, or substantially all, of the valve as a coherent structure. The system may then track the anatomical elements, for example using a predictive algorithm like a Kalman filter. The system may anticipate motion of each element due to heartbeat, patient breathing, and so on. The system may similarly maintain a stable lock on the element, for example via minor or micro adjustments to pose of the ICE catheter.

[0060] As one example, as new frames arrive, each active track may be predicted forward using a motion model that reflects typical intracardiac dynamics. Examples include a Kalman or related filter whose process model incorporates cardiac and respiratory cycles, optionally phase-aligned to ECG and respiration signals so that periodic motion is anticipated rather than chased. The system may then associatecurrent observations to predicted tracks using cost metrics such as centroid displacement, contour overlap, boundary-to-boundary distance, and feature similarity; data association may be solved, as one example, with a Hungarian assignment subject to gating thresholds that reject implausible matches. When the view changes abruptly or structures are partially occluded, dense optical flow or learned correspondence fields can propagate contours forward to bridge brief gaps, and tracks may persist through occlusions using a limited coast period with growing uncertainty until observations reappear.

[0061] Because ICE frequently presents partial sectors and rapid view shifts, the system may manage element identity carefully under splitting and merging. For example, a tracked wall segment that momentarily fragments under shadowing can be reconciled by continuity constraints and by preferring associations consistent with the prior model location; conversely, two adjacent structures that fuse in the image may be maintained as separate tracks using motion and shape cues. Track life-cycle logic handles birth, confirmation, dormancy during occlusion, and retirement after prolonged absence, while confidence from the perception model down-weights updates from ambiguous regions. When device elements are present, analogous tracks are maintained for device tips or shafts so that anatomy-device relationships are continuously available. The result is a set of stable, identity-preserving trajectories for key cardiac structures that remain usable despite partial visibility, providing reliable inputs for localization, proximity estimation, and closed-loop view stabilization.

[0062] The system may then analyze pose (e.g., position, orientation, and so on) of the anatomical elements within the field of view of the ICE catheter. For example, the system may compare the pose of the elements to the patient mesh described above. The patient mesh may indicate, for example, vertices associated with the anatomical landmarks. Thus, the system may ascertain, such as via triangulation, the pose of the ICE catheter (e.g., the pose of the tip). In some embodiments, the system may determine a six-degrees of freedom pose (e.g., x, y, z, along with rotational orientation such as roll, pitch, yaw) relative to the heart’s anatomy.

[0063] As one example, the system may solve for a pose that best aligns observed anatomical landmarks (e.g., the anatomical elements) with the heart model whileoptionally enforcing an intravascular tip prior. Sensor fusion can be employed to increase observability, for example by incorporating catheter inertial signals, controlhandle telemetry, and spike ranging that estimates tip-to-wall distance along the beam. Confidence-weighted optimization then yields a pose estimate with an associated uncertainty, supporting reliable view maintenance and safe motion planning despite the partial field of view typical of ICE.

[0064] The system may generate data in substantially real-time, which may be presented in a user interface and / or used in downstream processes. An example of such data is anatomical labels. For example, the labels may include a list of anatomical structures currently recognized in view (e.g., in the imaging plane or volume). Example labels may include mitral valve, aortic valve, left atrial appendage (LAA). The system may also generate the estimated tip position and orientation in the coordinate frame of the patient model. The system may also generate proximity metrics such as nearest-surface distance, rate of approach, and trend. For example, the system may generate the shortest distance from the catheter tip to the nearest recognized heart wall or critical structure (e.g., the system may generate or output that distance to posterial wall is 7mm). This may be used as a safety metric among other uses, such as situational awareness. The system may also generate confidence values for both recognition and localization. If the view is unclear or ambiguous, the system may reduce the values. These signals can be streamed to a controller for stabilization and to a user interface for situational awareness.

[0065] User interface integration may overlay semi-transparent labels on the live ultrasound, provide a miniature 3D model showing the estimated catheter pose, and display compact heads-up readouts for view name, proximity, and confidence. Visual and audible alerts can draw attention when proximity drops below a threshold or when localization confidence degrades, while simple controls allow the operator to freeze, resume, or override stabilization.

[0066] In some embodiments, a visualization interface (e.g., of the user interface) integrates real-time intracardiac ultrasound with model-derived context and safety information to assist interpretation and control. The interface may renderthe live imaging stream while concurrently presenting a compact set of computed metrics, includinganatomy identification, catheter pose estimates, proximity to nearby tissue or devices, and confidence indicators that reflect the reliability of current inferences. The presentation is configured to enhance situational awareness without obscuring clinically relevant grayscale detail and may be adapted to different workflows and user preferences.

[0067] The interface may provide a live anatomical overlay in which recognized cardiac structures are highlighted on the ultrasound image using semi-transparent masks and unobtrusive labels. The overlay can deform and move in synchrony with the image so that boundaries and annotations remain registered as the view changes. In some implementations, the overlay opacity or edge emphasis may be modulated by confidence so that well-supported regions are emphasized and uncertain regions are down-weighted or rendered differently, such as with a dashed contour, thereby conveying both interpretation and certainty at a glance.

[0068] A three-dimensional navigator may be presented, for example optionally in a dedicated panel of the user interface, that includes a graphical representation of the patient heart model together with a graphical representation of the catheter and its current field of view. For example, the system may store information reflecting a model of the catheter. As another example, another imaging modality (e.g., TEE) may obtain ultrasound information of the catheter. The navigator can update position and orientation in real time, providing spatial awareness of where the imaging tip resides within the heart, including anatomy outside the instantaneous imaging plane. In some configurations, the navigator may also depict keep-out zones, regions of interest, or the recent trajectory of the catheter to support repeatable navigation and quick return to standard views.

[0069] A heads-up display may present key metrics as concise text or iconography in a non-intrusive area of the screen. Examples include the name of the currently recognized view, the catheter tip position and orientation with respect to the model coordinate frame, nearest-surface distance and related proximity measures, and an overall confidence value for recognition and localization. The heads-up elements may be color-coded and trend-aware so that changes in proximity or confidence are visible without diverting attention from the primary image.

[0070] Visual and audible alerts may be issued when a safety or quality threshold is crossed. For instance, if the estimated distance to a wall or critical structure falls below a configurable limit, the proximity readout may change color and flash while a distinct tone is played to prompt timely attention. Additional alerts may be generated for loss of localization confidence or for entry into a keep-out region. In some embodiments, alerts can be paired with simple controls that allow the operator to acknowledge, pause, or override automation while maintaining a clear record of system state and rationale.

[0071] Figure 3 is a flowchart of an example stabilization process 300 for a robotically controlled intracardiac echocardiography (ICE) catheter. For convenience, the process 300 will be described as being performed by a system of one or more computers or processors (e.g., the system environment 100).

[0072] In the illustrated arrangement, process 300 may perform one or more of detecting unintended motion, consulting a patient-specific heart model, determining corrective control inputs, and / or executing closed-loop adjustments to restore a selected intracardiac view after a disturbance. In contrast to manual reacquisition, the disclosed technology enables rapid return-to-view using image- and sensor-informed control while respecting intravascular motion constraints and safety limits, such as proximity-aware keep-outs and view-based workflows maintained by the robotic system.

[0073] At block 302, the system obtains information indicating movement or disturbance of an ICE catheter (e.g., indicia of movement). In some embodiments, the system analyzes live, or otherwise substantially real-time, ultrasound information to detect rapid, nonphysiological view changes. For example, the system may leverage computer-vision techniques or cues such as optical flow or correspondence fields, while optionally cross-referencing simple physiological signals to discount expected periodic motion (e.g., cardiac motion, breathing, and so on). Based on detecting an abrupt deviation, the system infers a bump or nudge event and triggers stabilization, lin some embodiments, this may include returning to a previously viewed cardiac view (e.g., viewpoint of the ICE catheter).

[0074] Additionally or alternatively, the system may ingest sensor signals available on or near the catheter and robot, including accelerometers, gyroscopes, load or pressure sensors, and motor telemetry. As may be appreciated, these signals can revealtransient impacts, bending, or shaft torsion that may not yet be evident in the image stream, and can be sampled at high rate to capture sub-resolution vibrations or jolt signatures indicative of contact (e.g., disturbance). The disturbance evidence may further incorporate auxiliary device data, such as ventilator or ECG traces, so that the classifier distinguishes external perturbations from routine respiratory or cardiac motion prior to issuing corrective actions. Similar the above, with respect to at least Figure 3, the system may trigger the ICE catheter to return to a pre-disturbance viewpoint or pose.

[0075] At block 304, the system accesses a patient mesh to implement stabilization. In one configuration, the system retrieves a patient-specific mesh that has been generated or updated from intracardiac ultrasound during the procedure by deforming an average heart model, for example as described above. The patient mesh can be used to interpret the disturbed view by associating observed structures or anatomical elements (e.g., as described above) with known surfaces and landmarks. For example, and as described above, the system may perform triangulation to determine its current viewpoint (e.g., as compared to the prior viewpoint). In some embodiments, feature evidence is projected to mesh vertices and refined by a graph-model engine to keep the mesh current as new images arrive, which further improves view reacquisition after motion transients.

[0076] At block 306, the system determines adjustments to the pose of the ICE catheter to be performed via robotic control. The system identifies a target pose corresponding to the pre-disturbance view or a workflow-defined canonical view and computes micro-adjustments across available degrees of freedom, including axial translation, shaft rotation, and distal tip flexion. The resulting control vector may be derived from image-to-mesh correspondences and any sensor-detected motion and may be bounded by safety policies that maintain intravascular placement and enforce velocity and position limits near sensitive tissue. In implementations where the robot couples to the catheter handle dials, the control vector is translated into precise dial rotations and shaft manipulations by a drive mechanism while monitoring for backlash or slip, enabling fine-grained, repeatable changes that converge the current view back to the target view. When a disturbance exceeds a configured threshold, the controller mayescalate from incremental corrections to a structured view-reacquisition routine that uses the patient mesh to guide a short navigation back to the saved vantage point.

[0077] In some embodiments, the system may determine prioritization associated with degrees of freedom. For example, the system may prefer or prioritize rotation of the ultrasound transducer (or probe) about the longitudinal axis of the catheter body over flexion. In one example, flexion may be preferred over rotation of the tip portion when the tip portion is flexed relative to a more proximal portion of the catheter body. Rotation of the tip portion when flexed may be considered a less or least preferred motion and such motion may be implemented after other forms of adjustment of the transducer or probe are performed. However, this may be dependent on the type or capability of the ICE catheter. For example, the system may determine the catheter type and available degrees of freedom from device metadata or operator selection and apply modalityspecific constraints. For a design that supports decoupled distal tip rotation, the controller can prefer rotation-in-place about the imaging axis to reacquire orientation with minimal lateral sweep, followed by small flexion and translation to re-center the target and restore working distance. For a design in which shaft rotation under flexion produces a circular or orbital tip path, the controller may first reduce flexion toward neutral, execute the required shaft rotation to set azimuth, and then re-apply measured flexion to re-establish the view, thereby limiting unintended wall approach. Limits on allowable rotation while flexed, torque and rate bounds, and proximity-based keep-outs derived from the patient mesh may be enforced throughout, with backlash compensation and dwell timing applied to ensure repeatable, low-jitter corrections and to maintain standoff when proximity margins are small.

[0078] In some embodiments, the system may train a machine learning model, or access a trained model, to infer corrective motions from multimodal inputs that reflect image dynamics and / or device-state changes. The system may accumulate information including ultrasound information (e.g., from ICE catheter) together with catheter-borne sensor signals, such as pressure, accelerometer, and gyroscope readings from sensors, robotic device sensor telemetry 118, and additional data from other devices including, for example, ventilator or EKG outputs that correlate with physiologic motion and anticipated disturbances; these data streams are stored for model training and inferenceand are available to data analysis module 122, 148 to support determination of viewpreserving adjustments. During training, the model learns to classify disturbance type and to regress a continuous change-in-pose vector for the kinematics of the ICE probe based on cues such as rapid viewpoint shifts identified in the ultrasound stream, for example via optical flow, and high-rate sampling that reveals sub-resolution vibrations, with a patient-specific model such as patient mesh providing geometric context and distance-to-tissue constraints for stabilization and collision awareness outside the immediate field of view. Once trained, the adjustment model outputs a pose-delta that restores or maintains a preselected view following bumps, nudges, or workflow-driven transitions, with the system using the same, or similar, multimodal signals to continuously monitor for disturbances and to refine view-specific corrections over time.

[0079] At block 308, the system navigates, or otherwise causes adjustment of, the ICE catheter. The system, such as the robotic device, may execute the computed commands through the drive mechanism, then evaluate ultrasound information feedback to assess residual error and iteratively refines the pose until the target view is recovered within tolerance. For example, the system may determine a pose associated with the pre-disturbance viewpoint. In this example, the system may determine adjustments to degrees of freedom to cause a return to the pre-disturbance viewpoint, for example based on the above-described triangulation. For example, the system compare positions of landmarks as seen at the pre-disturbed and disturbed viewpoint. In some embodiments, the system may compute a matrix, pose-delta, mapping, and so on, between these viewpoints. The system may then trigger actuation of robotic control to adjust degrees of freedom of the ICE catheter.

[0080] In some embodiments, the system trains a machine-learning model that converts the desired pose-delta and current system state into executable instructions for drive mechanisms (e.g., mechanisms 120) to manipulate ICE catheter. In some embodiments, the system may train the model based on images, actuator states, command histories captured during acquisition, and / or maintenance of target views to learn an inverse-kinematics mapping for the ICE catheter that outputs bounded command increments foraxialtranslation, shaft rotation, and deflection, consistentwith the drive mechanism’s capability to actuate the catheter controls and reposition theimaging probe within the patient. At runtime, the system may generate instructions that maintain a preselected view or translate between views in an operator-specified workflow, with the patient mesh used to plan safe motion, enforce proximity constraints, and / or avoid occluded structures even when anatomy lies outside the current field of view. Additional device data, such as ventilator outputs, may be incorporated to anticipate physiologic motion and compensate proactively during the control rollout. Thus, block 308 may, in some embodiments, provides closed-loop robotic control that transforms learned pose adjustments into motor-executable instructions to efficiently stabilize the view and kinematically translate between clinically relevant ICE views under safety and workflow constraints.

[0081] While the above described use of a machine learning model, in some embodiments the system may navigate programmatically. For example, the system may use a kinematic model, optionally in combination with the graph model, a map associated with the catheters path through the heart, and so on, to determine navigational controls.

[0082] Additional techniques may be used and fall within the scope of the disclosure herein. For example, the system may determine a plane or volume through the patient mesh associated with the viewpoints. In this example, the system may determine a mapping or adjustment to cause the ICE catheter’s current view plane or volume to correspond with the pre-disturbance view plane or volume.

[0083] During this closed-loop execution, the system can maintain a running estimate of proximity to nearby tissue or devices and can trigger visual or audible alerts if a safety threshold is approached, while continuing to bias motion planning using the patient mesh so that corrections remain anatomically plausible and efficient. In some embodiments, the same control framework supports proactive stabilization during routine respiration and cardiac motion and can reduce overall navigation time by rapidly reestablishing standardized views selected in a user workflow once the catheter has been disturbed. Thus, process 300 restores the selected intracardiac view after bumps or nudges by combining disturbance detection, model-aware localization, and constrained robotic control grounded in a continuously updated patient mesh.

[0084] Additional example description related to robotically controlled ICE isincluded below. The resulting system, such as system 100, which includes or otherwise enables robotically controlled ICE may include one or more of the following features.

[0085] A non-robotic stand may, at least in part, holds an ICE catheter steady on a bed (e.g., no more wet towel required). For example, the stand may be coupled to the bed. The stand may releasably receive the proximal end of the ICE catheter and have one or more controllers to articulate the tip thereof. As described a robotic device 102 may hold the ICE catheter steady, for example to maintain a consistent view. The robotic device 102, for example with the console system 108, may make micro-adjustments to degrees of freedom to ensure the consistent view. The robotic device may have motion constraints once in the target anatomical area (e.g., Left atrium) that enables another operator (e.g., not including the interventional cardiologist) to safely move tip while manipulating the ultrasound image. The environment may have one or more of bed mounted, sterile connector, insertion end sheath stabilizers. The robotic controls enable enhanced stabilization as compared to current ICE systems.

[0086] Mechanically or robotically constrained to prevent pushing the ICE catheter too far. By limiting motion to prevent unwanted tissue contact, operators other than the interventional cardiologist can control ICE catheter, enabling a single interventional cardiologist to perform the image guided procedure. If the limits on movement are provided, a wider range of medical personnel is enabled to manipulate the ICE catheter once it’s in place (e.g., across the interatrial septum). Reaching a “safe zone” allows a handoff between operators, e.g., between an interventional cardiologist and another operator.

[0087] For example, catheter motion may be mechanically and / or robotically constrained to prevent excessive advancement and unwanted tissue contact. The system described herein, such as the environment 100, may enforce limits on translation, rotation, and distal tip flexion based on proximity sensing available on the catheter and / or positional awareness derived from ultrasound imaging registered to a patient mesh (e.g., heart model). By bounding motion within prescribed limits, operators other than the interventional cardiologist can manipulate the ICE catheter once positioned, enabling the cardiologist to perform the image-guided procedure while delegating view maintenance. As one example, after traversal of the interatrial septum, movementcontrol in the right and left atria may be performed with a higher degree of control than in esophageal contexts to maintain safe intra-chamber operation.

[0088] In some embodiments, a safe zone may be defined as a dynamically adjustable three-dimensional volume within the patient-specific heart model. An operator may establish the safe zone, for instance encompassing a target chamber after successful septal crossing, and the system confirms that the imaging tip is within this zone. Upon confirmation, a constrained mode of operation is activated in which the robotic drive mechanism accepts only commands that keep the catheter tip inside the safe zone and outside defined keep-out regions, thereby preventing accidental contact with critical structures and enabling a secondary operator to perform fine view adjustments. Commands that would move the tip outside the zone may be saturated, modified, or rejected according to safety policies.

[0089] The system may define, and enforce, the safe zone, in some embodiments, based on proximity sensing integrated with the catheter, spike-based ranging or other distance estimation, and / or imaging-based localization relative to the patient model. The safe-zone boundary can be updated in real time to account for cardiac and respiratory motion and may encode chamber-specific practices, such as tighter constraints in the left atrium. The system may provide an explicit handoff when the safe zone is reached, enabling an expanded range of medical personnel to manipulate the catheter under constraint, and may return control to the primary operator upon exiting the zone or upon override. In some embodiments, a graphical representation of the safe zone may be included in a user interface.

[0090] Thus, in some embodiments, safety controls can be based on proximity sensing built into the catheter or positional awareness based on imaging (e.g., internal and external imaging, such as ICE, X-Ray fluoroscopy). In some embodiments, movement control in the right atrium and left atrium may involve higher degree of control than in esophagus such as with TEE probe usage.

[0091] As described herein, the system may autonomously or semi-autonomously navigate the ICE catheter. For example, the system may adjust real-time position via sensing or detecting change in position and / or pose based on ultrasound information (e.g., images) and control the ICE catheter to maintain position and / or pose. The systemmay perform 3D mapping to target a catheter position based on ICE ultrasound information (e.g., images) and navigate to that position (e.g., Atrial septum crossing following interventional catheter, such as Watchman™ delivery system). This may include motion planning resulting in reduced manipulations, or steady positioning from Al that reduces manipulations.

[0092] With respect to Watchman™ system procedure, a cardiologist’s approach to crossing the interatrial septum may involve crossing with a sheath, then retracting the sheath while remembering where the sheath was in the imaging. Motion of the patient, image or gaze of the cardiologist may create uncertainty and add to procedure time. In contrast, the disclosed technology may enable use of an anatomical landmark / element map to inform where the crossing location is to facilitate moving, such as robotically controlled movement, the ICE catheter across the interatrial septum.

[0093] In some embodiments, the system may use image fusion to combine X-ray and / or TEE imaging with ICE imaging to help with crossing maneuver. In some embodiments, and as described herein, the system may perform stabilization techniques. For example, based on the if ICE catheter imaging probe being moved by another interventional device, the system can move the imaging probe back to prior position and / or pose. Al analysis of image and / or sensor outputs can be inputs to a robotic device (e.g., processor 110) and / or console system 108 to maintain position despite being moved or disrupted by the catheter. The sensor can detect movement or a load on the ICE catheter. Image analysis can indicate a direction of movement or need for repositioning.

[0094] Collision detection can be detected in the ultrasound information. For example, collision may use the above-described heart model. For example, the heart model may to determine or estimate distance between tissue and / or an interventional device. The heart model may provide for a complete, or substantially complete, view outside of the ultrasound’s field of view. For example, the graph model may represent a deformed average heart such that the graph model reflects the patient’s heart.

[0095] With respect to devices, an Al or classical technique may be used to detect jilt or bouncing off of something. For example, the technique may determine flutter or measures of flutter. In some embodiments, a mode may be used to detect position on aregular basis and can detect telltale signs that the catheter tip bumped into another object. The system may sample at a high rate enables to detect sub-resolution vibrations. The system may perform collision avoidance via a graph model. An example graph model may include a graph neural network, a graph convolution network, and so on. As described above, ICE can provide a partial view, and the graph model (e.g., heart model) can provide an approximation of the model outside of the field of view. This may enhance ICE because the operator may not have to turn the ICE catheter tip 180-degrees to see low interest areas (e.g., back side of the right atrium). Robotics can be used to move around without distracting the users to survey additional heart regions and / or record images that may indicate potential complications and then quickly return to a position providing guidance of a procedure.

[0096] Al-based image processing (e.g., image processing of ultrasound information from ultrasound sensor on ICE catheter, such as at tip of catheter). This may include example technical improvements. For example, recognition of structures and localization (e.g., provide information on questions such as: what is in the field of view? Where is the imaging device relative to the anatomy?). As another example, early recognition of complications such as pericardial effusion (system provides capability to “stare” (e.g., analyze ultrasound sensor data) substantially all the time, or interleaved with other imaging, or have mode / image cycling (e.g., depth, zoom, doppler flow). As anotherexample, automated measurements, device selection based on priors, ordevice assessment (anatomy measurements forevaluation, device selection, criteria for device such as PASS and CLOSE criteria for Watchman™ / Amulet™ devices). As another example, continually scanning the active image for visually detectable complications, e.g., pericardial effusion.

[0097] Robotically controlled ICE can optionally focus on valve or LAA (e.g., obtain view of valve or LAA using ultrasound sensor) and divert view for a threshold amount of time to compare the distance between the heart wall and the pericardial sac. If this distance is out of threshold or growing unacceptably the system can generate an alert to the interventionalist or operator. This effect would not, as one example, be the core focus of the interventionalist or operator, but the system can opportunistically look for it and give an alert that a complication may be happening. The system, such as Al, mayperform depth and zoom, turn doppler on and off, perform background monitoring. The system may collect information about what the catheter tip is looking at and where the catheter tip is located. Labeling may be included in a user interface of a display. Automated measurements may be performed, which may be useful for LAA device (e.g., Watchman or Amulet) sizing and outcome. Use of one slice, such as an ultrasound slice, only would make the detection of size hard to confirm. But ICE can do multiple slices to give 3D information. The graph model can provide some overlay information. The graph model may have a certainty / uncertainty capability. For example, the graph model may determine uncertainty associated with different portions of the graph model (e.g., vertices). In some embodiments, thresholding may be used to update presentation, or use, of the model. As an example, for portions below a threshold a wireframe may be presented while for portions above the threshold a smooth model may be presented.

[0098] An ICE catheter may optionally be modified, such as ICE catheter 104. For example, catheterwith tip sensingto replace physician tactile feedbackwhen navigating “free anatomic space.” As another example, the system may combine tactile feedback from catheter and visual feedback to help navigate. As another example, a pressure sensor may be integrated into the tip of the catheter, on the robotic device, and so on. In some embodiments, the pressure sensor may inform pressure being applied via the tip on the patient’s anatomy. It may represent a pressure associated with the catheter shaft, such as pressure applied to portions of the robotic device. ICE may be forward-facing and may use example feedback, such as resistance measurements. For example, an operator may understand that something feels unexpected. In this example, the system may use sensors to measure such interactions.

[0099] Fusion of multi-modal input may be used by the system. For example, fusion of ICE / X-Ray and / or ICE / TEE / CT. The system may perform 3D mapping to target a catheter position based on ICE-ultrasound information and then navigate to that position, using X-Ray or TEE for localization feedback. Image registration / guidance may be used. Pivoting / fulcrum effect may be used. The system may model along with knowledge of locations of portions of an ICE catheter tip and may provide information about possible location and shape of other portions of the ICE catheter, e.g., bending or snaking of portions.

[0100] A spike reading may be obtained from the ultrasound sensor. The system may perform distance calculations based on this reading or readings. This may represent a mode, which may optionally be interleaved with imaging mode(s). For example, the ultrasound transducer of the ICE catheter may be operated in a spike ranging mode which emits a brief pulse and measures the time of the first significant return to generate a spike reading indicative of distance to the nearest structure along the beam. The spike ranging mode may be time-interleaved with conventional imaging so that spike-derived distance estimates are updated at a higher rate than image frames and are used by the control system for stabilization, collision avoidance, and view reacquisition.

[0101] The system, such as the robotic device 102, may have a threshold number of degrees of freedom (e.g., 4, 6, and so on) and ICE may have additional degrees of freedom enabling ultrasound imagingto be registered with pre-procedural imagingof the patient (intraoperatively, for example).Other Embodiments

[0102] All of the processes described herein may be embodied in, and fully automated, via software code modules executed by a computing system that includes one or more computers or processors. The code modules may be stored in any type of non-transitory computer-readable medium or other computer storage device. Some or all the methods may be embodied in specialized computer hardware.

[0103] Many other variations than those described herein will be apparent from this disclosure. For example, depending on the embodiment, certain acts, events, or functions of any of the algorithms described herein can be performed in a different sequence or can be added, merged, or left out altogether (for example, not all described acts or events are necessary for the practice of the algorithms). Moreover, in certain embodiments, acts or events can be performed concurrently, for example, through multi-threaded processing, interrupt processing, or multiple processors or processor cores or on other parallel architectures, rather than sequentially. In addition, different tasks or processes can be performed by different machines and / or computing systems that can function together.

[0104] The various illustrative logical blocks, modules, and engines described in connection with the embodiments disclosed herein can be implemented or performed by a machine, such as a processing unit or processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A processor can be a microprocessor, but in the alternative, the processor can be a controller, microcontroller, or state machine, combinations of the same, or the like. A processor can include electrical circuitry configured to process computer-executable instructions. In another embodiment, a processor includes an FPGA or other programmable device that performs logic operations without processing computerexecutable instructions. A processor can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. Although described herein primarily with respect to digital technology, a processor may also include primarily analog components. For example, some or all of the signal processing algorithms described herein may be implemented in analog circuitry or mixed analog and digital circuitry. A computing environment can include any type of computer system, including, but not limited to, a computer system based on a microprocessor, a mainframe computer, a digital signal processor, a portable computing device, a device controller, or a computational engine within an appliance, to name a few.

[0105] Conditional language such as, among others, “can,” “could,” “might” or “may,” unless specifically stated otherwise, are understood within the context as used in general to convey that certain embodiments include, while other embodiments do not include, certain features, elements and / or steps. Thus, such conditional language is not generally intended to imply that features, elements and / or steps are in any way required for one or more embodiments or that one or more embodiments necessarily include logic for deciding, with or without user input or prompting, whether these features, elements and / or steps are included or are to be performed in any particular embodiment.

[0106] Disjunctive language such as the phrase “at least one of X, Y, or Z,” unless specifically stated otherwise, is understood with the context as used in general to present that an item, term, etc., may be either X, Y, or Z, or any combination thereof (for example, X, Y, and / or Z). Thus, such disjunctive language is not generally intended to, and should not, imply that certain embodiments require at least one of X, at least one of Y, or at least one of Z to each be present.

[0107] Any process descriptions, elements or blocks in the flow diagrams described herein and / or depicted in the attached figures should be understood as potentially representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or elements in the process. Alternate implementations are included within the scope of the embodiments described herein in which elements or functions may be deleted, executed out of order from that shown, or discussed, including substantially concurrently or in reverse order, depending on the functionality involved as would be understood by those skilled in the art.

[0108] Unless otherwise explicitly stated, articles such as “a” or “an” should generally be interpreted to include one or more described items. Accordingly, phrases such as “a device configured to” are intended to include one or more recited devices. Such one or more recited devices can also be collectively configured to carry out the stated recitations. For example, “a processorconfigured to carryout recitationsA, B and C” can include a first processor configured to carry out recitation A working in conjunction with a second processor configured to carry out recitations B and C.

[0109] It should be emphasized that many variations and modifications may be made to the above-described embodiments, the elements of which are to be understood as being among other acceptable examples. All such modifications and variations are intended to be included herein within the scope of this disclosure.

Claims

WHAT IS CLAIMED IS:

1. A robotically controlled intracardiac echocardiogram (ICE) system, the robotically controlled ICE system comprising: a robotic device comprising a housing comprising a recess having one or more drive mechanisms disposed therein; an ICE catheter comprising a catheter body having a proximal end and a distal end, an imaging probe disposed at a tip portion of the catheter body disposed adjacent to the distal end, and a control handle disposed at the proximal end, the control handle having one or more actuators configured to adjust one or more degrees of freedom of the tip portion of the ICE catheter; wherein the recess is configured to receive at least a portion of the control handle such that the one or more drive mechanisms engage the one or more actuators, wherein when engaged, the one or more drive mechanisms can actuate the one or more actuators to adjust one or more degrees of freedom of the tip portion of the ICE catheter; and one or more processors configured to execute instructions that cause the one or more processors to: analyze ultrasound information being obtained via the ICE catheter, the ultrasound information reflecting a viewpoint associated with an interior of a heart; detect indicia of movement of the ICE catheter; and cause implementation of one or more adjustments of individual degrees of freedom of the tip portion based on a patient mesh including a plurality of vertices forming a contour associated with the heart, wherein the adjustments are configured to update a pose associated with the ICE catheter to correspond with the viewpoint.

2. The robotically controlled ICE system of claim 1 , wherein detecting indicia of movement is based on the ultrasound information.

3. The robotically controlled ICE system of claim 1 , wherein detecting indicia of movement is based on sensors proximate to the robotic device.The robotically controlled ICE system of claim 3, wherein the sensor is disposed in the housing of the robotic device.

5. The robotically controlled ICE system of claim 3, wherein the sensor is disposed in the control handle of the robotic device.The robotically controlled ICE system of claim 3, wherein the sensor is disposed in the tip portion of the catheter body.

7. The robotically controlled ICE system of claim 6, wherein the sensor comprises a position sensor.

8. The robotically controlled ICE system of claim 6, wherein the sensor comprises a load sensor.

9. The robotically controlled ICE system of claim 1 , wherein to detect indicia of movement the one or more processors are configured to analyze optical flow associated with the ultrasound information.

10. The robotically controlled ICE system of claim 9, wherein the one or more processors filter and / or cross-reference physiological signals including one or more of cardiac motion or breathing.11 . The robotically controlled ICE system of claim 9, wherein the one or more processors are configured to compare optical flow analysis of motion of the tip portion with a commanded motion of the tip portion.

12. The robotically controlled ICE system of claim 11 , wherein the one or more processors is configured to reduce or halt motion if the assessed motion is less than the commanded motion.

13. The robotically controlled ICE system of claim 11 , wherein the one or more processors is configured to present a user interface indicating a lack of correspondence between the commanded motion and the assessed motion.

14. The robotically controlled ICE system of claim 1 , wherein to cause implementation of adjustments, the one or more processors are configured to: identify, based on the patient mesh, first positions of anatomical landmarks relative to a first pose associated with the viewpoint; and identify, based on the patient mesh, second positions of the anatomical landmarks relative to a second pose associated with a different viewpoint associated with the movement of the ICE catheter; and determine the one or more adjustments based, at least in part, on a comparison between the first positions and the second positions.

15. The robotically controlled ICE system of claim 14, wherein to determine the adjustments, the one or more processors are configured to prioritize a first degree of freedom over a second degree of freedom.

16. The robotically controlled ICE system of claim 15, wherein the one or more processors are configured to: identify a type associated with the ICE catheter, wherein the identification is based metadata or user input provided via an operator, and wherein the type informs the prioritization based on a capability of the ICE catheter.

17. The robotically controlled ICE system of claim 15, wherein the first degree of freedom is associated with rotation of the probe within the tip portion and wherein the second degree of freedom is associated with flexion of the tip portion relative to a portion of the catheter proximal of the tip portion.

18. The robotically controlled ICE system of claim 15, wherein the first degree of freedom is associated with flexing or unflexing the tip portion and wherein the second degree of freedom is associated with rotating a flexed tip portion.

19. The robotically controlled ICE system of claim 1 , wherein the one or more processors is configured to define a safe zone comprising a volume of permitted movement around the tip portion, the volume of permitted movement comprising a volume at least partially separated by a gap from anatomy disposed around the tip portion.

20. The robotically controlled ICE system of claim 19, wherein the one or more processors is configured to dynamically determine the safe zone based on a current position of the tip portion.21 . The robotically controlled ICE system of claim 20, wherein the safe zone comprises a volume disposed within a heart chamber within which the tip portion is spaced.

22. The robotically controlled ICE system of claim 1 , wherein the one or more processors are configured to iteratively update the patient mesh based on ultrasound information received from the ICE catheter.

23. The robotically controlled ICE system of claim 22, wherein iteratively updating the patient mesh includes deforming an average heart mesh associated with an average patient population or iteratively deforming the patient mesh.

24. The robotically controlled ICE system of claim 1 , wherein the one or more processors are configured to perform substantially real-time stabilization via adjustment of the tip portion.

25. The robotically controlled ICE system of claim 1 , wherein the one or more processors are further configured to: update a user interface to present a graphical representation of the patient mesh.

26. The robotically controlled ICE system of claim 25, wherein the user interface is configured to present one or more of a workflow indicating an order associated with viewpoints to be obtained via the ICE catheter, anatomical labels of anatomical landmarks depicted in the ultrasound information or on the patient mesh, proximity metrics, confidence values associated with the patient mesh, or a graphical representation of a safe zone.

27. The robotically controlled ICE system of claim 1 , wherein implementation of the one or more adjustments is based on inference of at least one machine learning model.

28. A method implemented bythe robotically controlled ICE system of any one of claims 1 -27.

29. A method implemented by a robotically controlled ICE system comprising one or more processors, the method comprising: analyzing ultrasound information being obtained via an ICE catheter of the robotically controlled ICE system, the ultrasound information reflecting a viewpoint associated with an interior of a heart; detecting indicia of movement of the ICE catheter; and causing implementation of one or more adjustments of individual degrees of freedom of a tip portion of the ICE catheter based on a patient mesh including a plurality of vertices forming a contour associated with the heart, wherein the adjustments are configured to update a pose associated with the ICE catheter to correspond with the viewpoint.

30. Non-transitory computer storage media storing instructions for execution by a robotically controlled ICE system, the robotically controlled ICE system comprising: a robotic device comprising a housing comprising a recess having one or more drive mechanisms disposed therein; an ICE catheter comprising a catheter body having a proximal end and a distal end, an imaging probe disposed at a tip portion of the catheter body disposed adjacent to the distal end, and a control handle disposed at the proximal end, the control handle having one or more actuators configured to adjust one or more degrees of freedom of the tip portion of the ICE catheter; wherein the recess is configured to receive at least a portion of the control handle such that the one or more drive mechanisms engage the one or more actuators, wherein when engaged, the one or more drive mechanisms can actuate the one or more actuators to adjust one or more degrees of freedom of the tip portion of the ICE catheter; and one or more processors configured to execute any one of claims 1 -27.

31. Non-transitory computer storage media storing instructions that when executed by a robotically controlled ICE system comprising one or more processors, cause the one or more processors to: analyze ultrasound information being obtained via an ICE catheter of the robotically controlled ICE system, the ultrasound information reflecting a viewpoint associated with an interior of a heart; detect indicia of movement of the ICE catheter; and cause implementation of one or more adjustments of individual degrees of freedom of a tip portion of the ICE catheter based on a patient mesh including a plurality of vertices forming a contour associated with the heart, wherein the adjustments are configured to update a pose associated with the ICE catheter to correspond with the viewpoint.

32. A robotic device, comprising: a housing comprising a recess having one or more drive mechanisms disposed therein, the recess configured to receive at least a portion of a control handle of an ICE catheter such that the one or more drive mechanisms engage the one or more actuators of the ICE catheter, wherein when engaged, the one or more drive mechanisms can actuate the one or more actuators to adjust one or more degrees of freedom of a tip portion of the ICE catheter; and one or more processors configured to execute instructions that cause the one or more processors to: analyze ultrasound information being obtained via the ICE catheter, the ultrasound information reflecting a viewpoint associated with an interior of a heart; detect indicia of movement of the ICE catheter; and cause movements of the one or more actuators to cause implementation of one or more adjustments of individual degrees of freedom of the tip portion.

33. The robotic device of Claim 32, wherein causing implementation of one or more adjustments is based on a patient mesh including a plurality of vertices forming a contour associated with the heart, wherein the adjustments are configured to update one or more degrees of freedom associated with the ICE catheterto correspond with the viewpoint.

34. A system comprising the robotic device of Claim 32, and an ICE catheter comprising a catheter body having a proximal end and a distal end, an imaging probe disposed at the tip portion of the catheter body disposed adjacent to the distal end, and a control handle disposed at the proximal end, the control handle having one or more actuators configured to adjust one or more degrees of freedom of the tip portion of the ICE catheter.

Images