System, method, and tool for visualization, segmentation and target selection from data acquired using multiple diagnostic methods

Abstract

The system described herein includes a tool that has been developed for the interactive planning of therapies (e.g., ablation therapies such as RA in the treatment of VT or catheter- based therapies, among many others) by accurately registering all treatment planning images into the same frame-of-reference, to enable physician interaction with different image visualizations of these registered images, including ablation target selection, to provide an environment that facilitates the efficient and accurate exchange of knowledge between physicians. The proposed tool is provided to increase access, expedite delivery, and improve target accuracy for various therapies, including the non-limiting example described below which provides an innovative non-invasive VT ablation therapy.

Description

SYSTEM, METHOD, AND TOOL FOR VISUALIZATION, SEGMENTATION AND TARGET SELECTION FROM DATA ACQUIRED USING MULTIPLE DIAGNOSTIC METHODSCROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 555,645 filed on February 20, 2024, the entire contents of which are incorporated herein by reference.TECHNICAL FIELD

[0002] The following generally relates to presenting data acquired using multiple diagnostic methods, in particular, to enabling visualization, segmentation, and target selection in multiple diagnostic methods, for example in presenting medical imaging data used in procedure planning.BACKGROUND

[0003] Ventricular tachycardia (VT) is considered the most common lethal arrythmia predominantly originating in the left ventricle (LV) [1], Current VT treatments including anti- arrhythmic drugs, implantable cardioverter defibrillators and catheter-based radio-frequency ablation are imperfect [2], Catheter ablation is not widely available; when performed it is expensive, invasive and long-term VT suppression is limited [3], Stereotactic external beam radiation (radio-ablation, RA) is an emerging non-invasive therapy which has shown promising early results [4], In treatment planning for RA, one or more diagnostic methods, including use of imaging modalities such as electroanatomic mapping (EAM) - including both invasive EAM and non-invasive EAM technologies such as electrocardiographic imaging (ECGi), contrast-enhanced cardiac-gated x-ray CT (CCT), magnetic resonance imaging (MRI), echocardiography (ECHO), single photon emission computed tomography (SPECT), and positron emission tomography (PET) are all used to help determine the underlying pathology, identifying the apparent origin of VT which is then targeted for RA. These technologies may also be used in planning other therapies, such as catheter-based ablation therapies (e.g., radiofrequency, cryotherapy, alcohol, and pulse-field ablations).

[0004] However, the multi-modality image integration required to plan and deliver RA is time consuming and complicated. This challenge extends to planning and treating other diseases. One or more of an electrophysiologist, a radiologist, and a radiation oncologist meet to review the medical images on stand-alone systems, and collaboratively identify, then select, the ablation target(s) on the radiation planning CT. Each clinician’s familiarity with different view planes and imaging modalities complicates the transfer of knowledge between these physician specialties. Defining a target on unregistered images relies on verbaldescriptors, and targets have been chosen historically by identifying discrete segments within the 17-segment heart model, which has been found to limit accuracy to approximately 5-7% of the LV myocardial surface area

[0020] ,

[0005] The treatment planning for RA thus requires information from multiple diagnostic methods, including imaging modalities, to determine the source of the VT and to develop an appropriate treatment plan. Different entities have reported different workflows and combinations of imaging including PET, SPECT, MRI, CT, invasive and non-invasive EAM, and ECG recordings [5], PET and SPECT are nuclear imaging modalities which can provide insight into the regional health of the heart muscle. MRI and CT, provide high-resolution information on the myocardium boundaries and the thickness of the myocardium, which can indicate regions of scar tissue. EAM and ECG provide information on the electrical activity of the heart, helping to identify where the arrhythmic heartbeat originates. Therapy planning may historically have been performed using CT, PET, 12-lead ECG and non-invasive EAM [6], Treatment planning, using multiple modalities and involving multiple clinicians, can create several challenges and consume valuable time.SUMMARY

[0006] In one aspect, there is provided a method of registering data of a first diagnostic to at least one other diagnostic, the method comprising: visualizing the first diagnostic relative to the at least one other diagnostic.

[0007] In certain example embodiments, registration of a non-computed tomography (CT) diagnostic is obtained by registering a CT scan that is registered to the non-CT diagnostic and applying a resulting transformation to the non-CT diagnostic.

[0008] In certain example embodiments, the non-CT diagnostic is positron emission tomography (PET) or single photon emission computer tomography (SPECT) and the preregistered CT scan is an attenuation correction CT.

[0009] In certain example embodiments, the non-CT diagnostic is electroanatomic mapping (EAM), and the pre-registered CT scan is the CT scan used in generating EAM geometry.

[0010] In certain example embodiments, the first diagnostic is EAM.

[0011] In certain example embodiments, the at least one other diagnostic is CT.

[0012] In another aspect, there is provided a method comprising: obtaining a 2D projection view of diagnostic scans; correlating the projection view to one or more 3D scenes; and displaying correlated data in a 3D imaging view.

[0013] In certain example embodiments, the projection view is a polar map.

[0014] In another aspect, there is provided a method comprising: using one or more segmentations from an imaging scan to sample the scan or other registered imaging scans within each segment; and using the sampling to generate one or more visualizations.

[0015] In certain example embodiments, the one or more visualizations comprises 3D objects or polar maps.

[0016] In certain example embodiments, the method includes using a segmentation from a CT scan to sample the other registered imaging scan, wherein the segmentation from the CT scan has been separated into a plurality of discrete segments.

[0017] In certain example embodiments, the other registered imaging scan comprises EAM, MRI, PET, or ultrasound.

[0018] In another aspect, there is provided a method comprising: selecting a volumetric target through a projection view by maintaining a correlation between a 3D scene to the projected view to maintain a dimension from the projection.

[0019] In certain example embodiments, the projection view is obtained using sampled information from one or more scan types.

[0020] In certain example embodiments, the one or more scan types comprise EAM, MRI, PET, or ultrasound.

[0021] In certain example embodiments, the volumetric target is delineated in the projection view and is correlated to a resulting transmural target by a preserved mapping.

[0022] In another aspect, there is provided a method comprising: selecting one or more targets in any one of a plurality of visualizations; and having the target update in other ones of the plurality of visualizations to show a corresponding selected region / volume.

[0023] In certain example embodiments, the target comprises an ablation target selected in one view and a corresponding target is displayed in the plurality of other visualizations.

[0024] In certain example embodiments, the ablation target is selected in a polar map and projected back to a 3D scene.

[0025] In certain example embodiments, a selected target is motion-tracked throughout a cardiac cycle to identify a summative volume occupied by the selected target throughout the cardiac cycle.

[0026] In certain example embodiments, any of the above methods may include providing an ability to assess in anatomic and cardiac slice views.

[0027] In certain example embodiments, any of the above methods may include providing an ability to export or integrate a selected target volume into a planning or treatment system.

[0028] In certain example embodiments, two temporal diagnostics are registered temporally in a cardiac cycle.

[0029] In another aspect, there is provided a method comprising: segmenting a myocardium across a plurality of phases of a cardiac cycle to identify at least one of wall thickness, wall thickening, and wall motion, with a segmentation that has been separated into a plurality of discrete segments.

[0030] In certain example embodiments, the plurality of discrete segments is obtained using a segmentation from a CT scan to sample the other registered imaging scan, wherein the CT scan has been separated into the plurality of discrete segments.

[0031] In another aspect, there is provided a computer system comprising: a processor; and a memory, the memory storing processor executable instructions that, when executed by the processor, cause the computer system to perform the method of any one of the above method and / or example embodiments.

[0032] In another aspect, there is provided a computer-readable medium storing processor executable instructions that, when executed by a processor of a computer system, cause the computer system to perform the method of any one of the above methods and / or example embodiments.BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Embodiments will now be described with reference to the appended drawings wherein:

[0034] FIG. 1a is a block diagram of an example of a computing environment, including a visualization tool.

[0035] FIG. 1 b is a block diagram of an example of a computing environment and the visualization tool configured to process ECG, EAM, CT and PET imaging modalities.

[0036] FIG. 2 is a schematic block diagram of the visualization tool.

[0037] FIG. 3 is a block diagram of a computing device.

[0038] FIG. 4 is a graphical user interface (GUI) rendered by the visualization tool.

[0039] FIG. 5 illustrates an EAM visualization.

[0040] FIG. 6 illustrates an example lead placement for ECG sampling.

[0041] FIG. 7 is a screen shot illustrating an ECG widget for temporal EAM playback.

[0042] FIG. 8 illustrates an axial slice of ECGi CT (A), PET CT (B) and co-registered CT scans (C).

[0043] FIGS. 9 and 10 illustrate an example segmentation wedge, in this example showing a 10-degree wedge and semi-transparent myocardium segmentation (A) and resulting intersection (B).

[0044] FIG. 11 illustrates an example of vertex notation for wedge coordinates used in creating discrete myocardium segments.

[0045] FIGS. 12A and 12B are PET imaging polar maps with (right) and without (left) 17-segment heart model indices.

[0046] FIGS. 13A and 13B provide an example 2D slice-view of CECT with PET myocardium object overlayed.

[0047] FIGS. 14 and 15 is an example 2D slice-view of PET and CECT alpha-blend.

[0048] FIGS. 16A, 16B, 16C, 16D, and 16E illustrate examples of data preparation workflows used to create multi-diagnostic visualizations.

[0049] FIGS. 17-22 are screen shots of example stages in utilizing the visualization tool to perform a planning and target selection workflow.

[0050] FIGS. 23A and 23B illustrate how wall motion and wall thickness calculations may be measured.

[0051] FIG. 24 shows a wall thickness, thickening and motion measurement on a shortaxis MRI slice.

[0052] FIG. 25 illustrates example data for possible use of RA in ejection fraction and heart failure.

[0053] FIG. 26 illustrates case planning duration data in an example study.DETAILED DESCRIPTION

[0054] The system described herein includes a tool that has been developed for the interactive planning of therapies (e.g., ablation therapies such as RA in the treatment of VT,atrial fibrillation and heart failure, or catheter-based therapies, among many others) by accurately registering all treatment planning images into the same frame-of-reference, to enable physician interaction with different image visualizations of these registered images (e.g., including ablation target selection in the RA example), to provide an environment that facilitates the efficient and accurate exchange of knowledge between physicians. The proposed tool is provided to increase access, expedite delivery, and improve target accuracy for various therapies, including the non-limiting example described below which provides an innovative non-invasive VT ablation therapy. While examples provided below may relate specifically to RA in the treatment of VT, it can be appreciated that the principles of the platform and tool described herein can be applied to other applications in which diagnostic data is acquired using multiple methods (e.g., including use of multiple imaging modalities) is to be visualized, segmented, and targets selected for subsequent planning / analysis, e.g., radiotherapy for heart failure, atrial fibrillation, or as an input to catheter ablation planning, etc.

[0055] While examples herein refer to visualization of the LV, the principles and tools described herein may be applicable to targeting other anatomic regions such as the right ventricle (RV) or atria. For example, in general, the following system may use polar maps or other projection images to present one or several anatomic regions in one or several 2D projections, and store values to enable the system to map from the projection space to 3D, and vice versa.

[0056] To that end, a software platform and workflow have been developed to facilitate precision-targeted therapy planning, including multi-diagnostic image registration (e.g., rigid and / or affine non-rigid) and 2D-3D-4D visualization across modalities. For example, electrocardiographic maps of VT parameters can be registered temporally to surface electrode data to recreate familiar ECG tracings. 2D polar maps, 3D slice-views, and 4D cine-renderings may be used for hybrid fusion displays of molecular and electroanatomic images. Segmentations of the cardiac-gated contrast CT blood-pool and molecular images of perfusion and glucose metabolism can be used to identify regions of fibrotic scar tissue and hibernating myocardium in the 3D scene. Ablation targets may then be “painted” onto a 2D polar map, 3D slice or 4D-cine view, and exported as DICOM for import to radiotherapy planning software. The combination of accurate multi-diagnostic image registration and visualizations may enable more reliable therapy planning, expedite treatment and may improve understanding of the underlying pathophysiology of these lethal arrhythmias.

[0057] The system, methods, and tool described herein have been developed recognizing certain challenges in current clinical workflows.

[0058] A first challenge relates to ECGi interaction in that current ECGi visualization software is only available on the ECGi acquisition system, which is not available during case-planning sessions. Currently, ECGi screen recordings and screenshots are brought to the case planning sessions, but they are not interactive, which limits their demonstration and interpretability by other clinicians.

[0059] A second challenge relates to interpretability of ECGi by nonelectrophysiologists in that the electrophysiologists may have the only expertise to interpret the ECGi maps, and other clinicians would then need to rely on verbal descriptors from the electrophysiologist to understand the regions of interest identified through ECGi.

[0060] A third challenge relates to the lack of data integration in that each imaging diagnostic method used in therapy planning is reviewed through a different software tool, often on different computers or monitors. However, the clinicians need to collaboratively identify, and select the desired ablation target region on the treatment planning CT scan, which can be complicated while using multiple sources and be prone to error.

[0061] A fourth challenge relates to the different visualization familiarity found between clinicians in that each clinician has experience working with different imaging modalities and different visualizations for interpreting the images. Electrophysiologists most commonly evaluate cardiac structures in 3D views. Imaging cardiologists often work with 3D, 2D slice views in cardiac and global anatomic planes, and polar map views. Radiation therapy clinicians are most frequently working with 2D slice views in anatomic planes. Collaboration between these clinicians is challenging when each is unfamiliar with their colleagues’ preferred visualizations.

[0062] In general, imaging cardiologists have the most diverse imaging skillset and may act as a translator between the other clinicians, however, the flow of information between clinicians and their respective imaging specialties remains a challenge. Most commonly, the 17-segment heart model acts as an intermediary between the electrophysiologists and the imaging cardiologists, as it’s a fra me- of- reference with which most cardiology clinicians are familiar, but this results in low resolution.

[0063] To address these challenges, the present system provides EAM and / or ECGi compatibility to address a lack of available tools. The system is thus configured to load, display, and enable interactions with electroanatomic maps exported from proprietary or commercially available EAM and / or ECGi mapping software. The EAM / ECGi visualization includes geometry information from the epicardial segmentation performed in the mapping software, prior to map export. The software tool may then also be compatible with thefollowing four map types: potentials, activation, directional activation, and propagation. In this way the user can assess all aspects of the ventricular epicardial surface with interactions, such as zoom, pan, and rotation in a 3D view. Playback of the temporal dimension of the potentials and propagation maps may also be provided and the software can load and display single-lead skin-surface measurements from the ECGi electrodes, exported from the mapping software, and playback of the maps can be temporally synched to these tracings.

[0064] The system is also configured to enable multi-diagnostic type image registration as it has been recognized that the current planning process may lack the ability to correlate information across imaging acquisition types and / or modalities. Image registration provides a mechanism to more accurately correlate spatial information between different imaging scans. To that end, the system enables multimodal image registration including the registration of, in the example described herein, ECGi to PET and CT scans. A diagnostic CT can be selected as the fixed image in image registration. Multiple CTs may be coregistered, but the same diagnostic CT can act as the fixed image for all image registrations. An automatic image registration method can be used but manual adjustments to the registered images should be available. Registrations may be performed to one or several phases of a temporal CT, as selected by the user.

[0065] The system described herein also provides a cohesive planning tool, to provide the ability to understand the features of interest from each of the imaging modalities used in the therapy planning of RA. To that end, the software provides a mechanism to identify the morphology of the VT and its relation to the cardiac anatomy, including information from PET, CT, and invasive or non-invasive EAM (e.g., ECGi) playback of the temporal data from a cardiac-gated CT scan may be made available and the visualization of the images can be interpretable by clinicians with differing expertise.

[0066] To address the above challenges, the system described provides ablation targeting. It has further been recognized that a challenge in the current therapy planning process is defining the target slice-by-slice on the radiotherapy planning CT. Once selected, there still may be a lack of confidence in the selected target in relation to other imaging modalities. This is not only time-consuming but increases the risk of error. To address this, the system provides a mechanism to accelerate the selection of a target for non-invasive RA, displaying the selected target relative to each imaging-related diagnostic method, and an ability to export or directly integrate this target to / into radiation therapy planning software.

[0067] The system also provides the ability to achieve shared decision making such that the collaboration between clinicians with different expertise for identifying an ablationtarget can be achieved, which may introduce challenges in communication of information across visualizations. To that end, the system should promote shared decision-making between clinicians of differing specialties by providing visualizations and interactions that meet the breadth of experience of the clinicians involved in the therapy planning process. The software described herein enables assessment of the images in 3D views, 2D-slice views including anatomic (axial, coronal and sagittal) and cardiac (LV short-axis and long- axis) planes, and polar map views. The software may promote the ability to correlate regions of interest between visualizations and imaging modalities.

[0068] Additionally, the system provides familiarity in that the clinicians involved in therapy planning have a busy schedule, and familiar software tools and interactions are more likely to receive clinical uptake and reduce the risk of misinterpretation when minimal training is necessary. As such, the system builds on familiar visualizations for medical imaging to promote rapid uptake and confidence with clinicians.

[0069] Referring now to the figures, FIG. 1a illustrates a computing environment 10 that spans both a clinical environment (e.g., when considering the equipment and clinical environment(s) required to obtain suitable imaging data) and a computer network or backbone that enables storage of data, usage of a visualization tool 12, and access to graphical user interfaces generated and rendered by the tool 12 by multiple different users 28, also referred to herein as clinicians, physicians, technicians, professionals, specialists, etc. It can be appreciated that the visualization tool 12 as described herein may include proprietary software programs, open-source programs, commercially available software programs, or a combination / bundle of both proprietary, open-source, and commercially available software programs, together implementing the visualization tool 12.

[0070] In the example environment 10 shown in FIG. 1a, multiple diagnostic techniques, medical images, processes or methods (referred to herein as “diagnostics 18”) that may be physiological, anatomical or other, are applied to a patient or subject 14. Such diagnostics 18 may include various imaging modalities, biosignals, etc. Each diagnostic 18 is acquired using acquisition system, device or generally any “apparatus 16”, which can range from a simple imaging device such as a camera, a biosignal interface, to a complex medical imaging machine such as an MRI machine. The acquisition apparatuses 16 are shown in dashed lines to generally denote an interface or interaction with the subject 14 by the respective diagnostic 18. In this example, an arbitrary N different diagnostics are applied, namely Diagnostic 1 , Diagnostic 2,..., Diagnostic N, each being referred to generally by numeral 18. Each diagnostic 18 generates acquired data, which may generally be referred to herein as “imaging” data 20, in this example Di, D2,..., DN. The sets of imaging data 20are collected by a data collection system 22 and stored in an imaging database 24. It can be appreciated that the data collection system 22 and imaging database 24 are shown for illustrative purposes to reflect any computing interface, module, function, application, or structure that is configured to obtain imaging data 20 for the purpose of subsequent use, e.g., by storing such imaging data 20 in the imaging database 24. As such, other configurations are permissible without departing from the principles discussed herein.

[0071] The visualization tool 12 is coupled to the imaging database 24 to access imaging data 20 for a corresponding subject 14 in order to conduct an analysis such as in treatment planning according to the examples herein. The visualization tool 12 may be accessed or run from a computing station 26 operated by a user 28 such as a first clinician or specialist. The visualization tool 12 is advantageously also coupled to a network 32 such as an internal hospital or clinic intranet or to a cloud-based or otherwise external system used by multiple users 28 from multiple corresponding external workstations 30. It can be appreciated that any networked or connected environment may be utilized such that multiple users 28 can access the visualization tool 12 to collaborate on an analysis or planning process. However, the same collaboration may also be performed using the same workstation 26 coupled directly to / with the visualization tool 12 should the application and environment permit.

[0072] The visualization tool 12 may also be coupled to or otherwise be part of a broader suite of applications, services, or tools within the computing environment 10. For example, as shown in FIG. 1a, the visualization tool 12 may be used to generate an output for a planning tool 34, which in this case may be accessible via the network 32 or coupled directly to or be within the same program as the visualization tool 12 (as illustrated optionally using dashed lines).

[0073] While the planning tool 34 is shown as being separate, to illustrate that the visualization tool 12 may be used in a workflow having downstream or subsequent stages, the visualization tool 12 may be a standalone program or service wherein any output or result may be used directly, e.g., within a clinical environment.

[0074] FIG. 1 b illustrates a specific implementation of the computing environment 10 for interactive planning of therapy such as RA in the treatment of VT as described above. In this example, four diagnostics 18 and corresponding imaging data 20 are utilized, namely ECG, EAM, CT, and PET; each utilizing a corresponding acquisition apparatus 16. Like elements in FIG. 1 b are denoted by like numerals, when compared to FIG. 1a, and details thereof need not be reiterated. However, in the example shown in FIG. 1 b, a Radioablation planningtool 134 is shown as part of a downstream stage that uses an output of the visualization tool 12, described in greater detail below.

[0075] While certain diagnostics 18 are shown by way of example in FIG. 1 b, it can be appreciated that various other diagnostics 18 in various other combinations are permissible. Table 1 below provides details concerning a sample of such diagnostics 18, including those shown in FIG. 1 b.Table 1: Diagnostic Method Examples

[0076] FIG. 2 illustrates a schematic functional block diagram of the visualization tool 12. The visualization tool 12 includes one or more database interfaces 40, such as APIs and the like, to enable the visualization tool 12 to obtain imaging data 20 which corresponds to a particular subject 14 being analyzed or evaluated. The visualization tool 12 also includes one or more computing interfaces 42 to permit the visualization tool 12 to be used by a computing terminal 26, 30 or other computing device 70 (e.g., as shown in FIG. 3 described below). The visualization tool 12 may be embodied as a software program or application, which may be deployed as a desktop application, a mobile application or both. The visualization tool 12 also includes one or more network interfaces 44, e.g., to enable it to communicate with and via the network 32 and / or other short- or long-range communication connections (e.g., WiFi, Bluetooth, IR, etc.) via the computing terminal 26, 30 or computing device 70 used to interact with the visualization tool 12.

[0077] The visualization tool 12 may be configured or programmed to include certain functionality as exemplified herein. The functional elements shown in FIG. 2 are for illustrative purposes and various other widgets, modules, functions, tabs, windows, features, and sub-programs can be integrated within the visualization tool 12 according to standard programming techniques. In this example, the visualization tool 12 includes an image registration and display module 46 configured to perform image registration amongst the imaging data 20 and render and display imaging data 20 within a GUI. Also shown is an image segmentation module 48, which may be configured to permit image segmentation operations as described further below. A target selection module 50 is also shown, which may be configured to enable target selection within the different image modality visualizations, also described in greater detail below. These modules 46, 48, 50 may interact with each other and / or with the interfaces 40, 42, 44 to enable the visualization tool 12 to be used in planning analysis operations. The target selection module 50 is shown as including an export function 52, to export data to a subsequent stage such as the planning tool 34. However, it can be appreciated that the export function 52 may be embodied as its own function or within a different module in other implementations.

[0078] FIG. 3 illustrates a schematic block diagram of a computing device 60, which may generally represent any user or clinician device operating within the computing environment 10. For example, the computing device 60 may generally represent a smart phone, tablet, desktop computer, server computer, virtual reality (VR) headset, augmented reality (AR) enabled device, controller, embedded device, etc.; and be adapted for the corresponding role of the entity within the computing environment 10. In this example, the computing device 60 includes one or more processors 62 (e.g., a microprocessor, microcontroller, embedded processor, digital signal processor (DSP), central processing unit (CPU), media processor, graphics processing unit (GPU) or other hardware-based processing units) and one or more network interfaces 64 (e.g., a wired or wireless transceiver device connectable to a network via a communication connection). Examples of such communication connections can include wired connections such as twisted pair, coaxial, Ethernet, fiber optic, etc. and / or wireless connections such as LAN, WAN, PAN and / or via short-range communications protocols such as Bluetooth, WiFi, NFC, IR, etc.

[0079] The computing device 60 also includes an application 72 (e.g., visualization tool 12, planning tool 34, etc.), a data store 74, and application data 76. The data store 74 may represent a database or library or other computer-readable medium configured to store data and permit retrieval of data by the computing device 60. The data store 74 may be read-only or may permit modifications to the data. The data store 74 may also store both read-onlyand write accessible data in the same memory allocation. In this example, the data store 74 stores the application data 76 for the application 72 that is configured to be executed by the computing device 60 for a particular role or purpose.

[0080] While not delineated in FIG. 3, the computing device 60 includes at least one memory or memory device that can include a tangible and non-transitory computer-readable medium having stored therein computer programs, sets of instructions, code, or data to be executed by processor(s) 62. The processor(s) 62 and network interface(s) 64 are connected to each other via a data bus or other communication backbone to enable components of the computing device 60 to operate together as described herein. FIG. 3 illustrates examples of modules and applications stored in memory on the computing device 60 and executed by the processor(s) 62.

[0081] It can be appreciated that any of the modules and applications shown in FIG. 3 may be hosted externally and be available to the computing device 60, e.g., via a network interface 64. The data store 74 in this example stores, among other things, the application data 76 that can be accessed and utilized by the application 72. The data store 74 may additionally store one or more software routines in a cache or in other types of memory.

[0082] As shown in FIG. 3, the computing device 60 may, optionally (e.g., when configured as a user device such as a smart phone, tablet, desktop computer, etc.), include a display 66 and one or more input device(s) 68 that may be utilized via an input / output (I / O) module 70. That is, such components may be omitted when the computing device 60 does not interact with a user.

[0083] FIG. 4 illustrates a software GUI for heart arrhythmia ablation planning. Portion A of the GUI shows a 3-lead ECG trace for playback of temporal EAM. Portion B of the GUI shows a set of target volume statistics. The graphs in Portion C are PET polar map views. It can be appreciated that polar maps may also be used to display other imaging modalities or acquisitions 18 relevant to planning, such as SPECT, MRI, CT, etc. Portion D of the GUI illustrates that, in this example, corresponding 4D-cine views with EAM (left) and PET (right) are available and provided. It can be appreciated that 4D cine views are optional such that if a non-dynamic image is uploaded, the visualization tool 12 can accommodate same. Portion E illustrates the 3D CT slice views with PET overlay. The blackened region on the polar maps, 4D and 3D slice views denotes the volume selected for ablation. FIG. 5 illustrates that the GUI shown in FIG. 4 may accommodate EAM or any other cardiac mapping diagnostic, electrical mapping or other. While the portions shown in FIG. 4 are illustrated together in a single view, it can be appreciated that the portions may be presented separately with anability to toggle between such views. The following discussion provides an example of how to process the imaging data 20 to generate the visualizations, to register and segment images, and to permit target selection.EAM Visualization

[0084] EAM is an important tool in therapy planning of non-invasive cardiac RA, offering insight into the source and morphology of arrhythmias. EAM can be generated invasively with a mapping catheter system or non-invasively with ECGi. One of the benefits of non- invasive cardiac RA is the non-invasive nature of the therapy, and so ECGi may be desirable in some cases, to maintain non-invasive imaging for treatment planning as well.

[0085] Currently, for the use of ECGi in therapy planning, there exists an inability to co-register the information with other diagnostics 18. The electrophysiologist can often pinpoint the underlying morphology and breakout location of the VT through an ECGi map, but correlating that location from ECGi to a therapy planning CT is challenging.

[0086] The visualization tool 12 has been configured to address this shortcoming by providing a function to load the EAM and look for a surface denoting endocardial or epicardial surface (or other), defined by vertices, point clouds, STL, etc. The visualization tool 12 may also be configured to look for a list of values corresponding to regions of this surface (e.g. the scalar values being encoded on the surface).EAM Map Types

[0087] From the body surface electrical potentials captured for a single heartbeat, various epicardial maps can be reconstructed. Each map provides information relating to the propagation of electrical activity across the heart, and each may offer different insights into the underlying morphology of the heartbeat, which is most useful in identifying the site of origin and direction of propagation of an arrhythmia.

[0088] For example, where s is either an epicardial or endocardial surface, t is time, and the system provides A(s) encoding any electroanatomic map, where A is the map value at the point s on the surface. This can be extended to A(s,t) for EAM with a temporal dimension.ECG Data

[0089] ECGi is a new imaging modality and clinical uptake to date has been limited. For this reason, few clinicians are experienced in its interpretation. ECG on the other hand is routinely used and is typically an electrophysiologist’s first tool in diagnosing VT [7], The maps are often used to identify the region of epicardium with the earliest activation of the VT beat, and the corresponding direction of activation from that point, and identifying the correct point in the ventricular tachycardia beat is important. The visualization tool 12 can implement an ECG-synchronized playback of the temporal ECGi maps to enable understanding of the point in time being displayed, as most clinicians are familiar with review of ECG tracings. To achieve this, ECG can either be imported from an ECG recording or simulated (or inferred) from similar body surface electrodes (e.g., from ECGi acquisition electrodes). For example, a 3-lead ECG has electrodes placed on the upper right chest (RA lead), upper left chest (LA lead), and lower-left torso (LL lead) as shown in FIG. 6. Similarly positioned ECGi acquisition electrodes can be used to simulate a 3-lead ECG. In the event that the electrode provides a low-quality signal, a dropdown list of alternative electrodes may be provided to change the electrode being used for each lead representation. In the graph (e.g., Portion A in FIG. 4 shown enlarged in the form of a widget in FIG. 7), a vertical line is used to denote the timepoint of the ECGi map displayed relative to the ECG beat. As shown in FIG. 7, an interactive handle can be provided to allow the user to scroll temporally through the ECGi map to a specific point in time, or the line plays along with the ECGi map when the play button is used, at a user-defined frame- rate. Frame-by-frame skipping forward and backward through the cycle is provided for fine temporal adjustments.Image Registration

[0090] Image registration is the process of aligning images of the same object obtained by multiple different imaging modalities and allows information to be accurately translated between different imaging modalities. The contrast-enhanced cardiac CT may be chosen as the fixed image for all image registrations. This CT is the highest resolution image acquired and is ECG-gated to the cardiac cycle offering good definition of the cardiac structures. The visualization tool 12 may be configured to permit the user to observe the registration result and make manual adjustments (if desired) - e.g., scaling, translation, rotation. This may be done to alter the resulting transformations or to perform them completely manually from scratch.ECGi to Cardiac CT Registration

[0091] Intramodality and intermodality image registration have been extensively studied for conventional imaging modalities, such as PET, CT, and MRI; however, no tools were publicly available for registering ECGi to other imaging modalities. There are numerous strategies for medical image registration, including manual, landmarkbased, surface, or volume-based registration. In addition, there are different constraints and loss functions that can be used to shape the result of the transformation. To perform this registration, the format of the exported ECGi data can be changed such that it contains the position of the ECGi mesh relative to the ECGi CT coordinates. From this, one can leverage the registration of the ECGi CT to a contrast- enhanced cardiac CT and apply the same registration transformation to the ECGi objects. A commercially available toolkit, known as SlicerElastix® may be utilized. These transforms can then be applied to other volumes or objects. These output transforms can be applied to the ECGi objects to register the ECGi maps to the cardiac CT. Within the SlicerElastix® toolkit, there are various preset registration parameters for different imaging modalities. The default registration parameters may be used with sufficient accuracy, initially based on visual assessment and later quantified, however, such registration parameters may also be refined and optimized. These default image registration methods use the Mattes mutual information algorithm [8], as is recommended for use with Elastix® for both intermodality and intramodality registrations, even though intensity-based registration methods are more commonly used in intramodality image registration [9], Table 2 below provides a list of registration parameters that may be used in the registration process.Table 2 - Registration Parameters for Image Registration Process

[0092] As a subjective visual assessment of the success of the image registrations, image fusion of the fixed and registered CT scan was assessed, considering various cardiac structures and their positional alignment postregistration, as shown in FIG. 8.Invasive EAM to Cardiac CT Registration

[0093] In the case of invasive EAM, there is unlikely to be a CT scan in the EAM frame-of-reference. If an EAM map covering a sufficient extent of the LV endocardium has been acquired, registration may be performed by surface-based registration, registering the EAM to the endocardial surface of the Cardiac CT myocardium segmentation. If aspects of the LV are not well mapped, manual or landmark-based registration methods may be preferred for registration

[0021] ,PET to Cardiac CT Registration

[0094] The approach developed to register PET images to the cardiac CT scans was similar to that described for ECGi. The requirements of reducing user interaction removed manual and landmark-based registration from consideration. Inter-modality image registration is often achieved with a mutual information similarity measure [9], and thus the 3D Slicer mutual information registration tools were first tested. Initial results of using mutual information to register PET volumes to the cardiac CT, though only visually assessed, were not considered satisfactory. Because our PET datasets are acquired with a PET-CT hybrid scanner, one may have attenuationcorrection CT (ACCT) scans pre-registered with the PET volumes. Though these ACCT scans have lower resolution, the feasibility of leveraging the ACCTs as an intermediary to register the PET volumes to cardiac CT scans was explored, similar to the method used for ECGi. As with ECGi registration, the SlicerElastix® toolbox may be used, wherein the attenuation CT can be cropped to a similar FOV as the cardiac CT, and rigid and non- rigid registrations evaluated using the same parameter set. The corresponding output transforms can be applied to the PET image volumes. When both FDG and Perfusion PET scans are performed, separate ACCT scans can be loaded and registered for each.MRI to CT Registration

[0095] MRI can provide useful information relating to the scar through late gadolinium enhancement (LGE) MRI. These images, when available, are also coregistered into the scene. This can be achieved through landmark registration, manual registration, or by reconstructing a 3D MRI from 2D slices, and leveraging mutual information or surface based registration methods, similar to those of the CT- to-CT registration using the SlicerElastix toolbox or previous registrations.CT Segmentation

[0096] In cardiac ablations, the usual intent is to ablate the full transmural thickness of the ventricular myocardium

[0010] , A method has been developed to generate a CT segmentation of the left ventricular myocardium to facilitate target selection.

[0097] There are many approaches to obtaining segmentations from medical images, including manual segmentation tools, semi-automatic or fully automated segmentation such as atlas-based or machine learning segmentation. The visualization tool 12 can accommodate these different segmentation methods, where different segmentation strategies may be applicable based on imaging artefacts, scan quality, and user experience.Creating Discrete Myocardium Segments

[0098] The resulting full left ventricular myocardium segmentation was divided into 576 discrete segments to facilitate target selection, and later to be used for PET visualization (described below). This division was achieved by dividing the total LV volume into 16 short-axis slices and 36 rotational slices about the LV long-axis.

[0099] User-defined position of the mitral valve center and LV apex may be used to define the LV long-axis, used for orienting the discretization of the myocardium segmentation. This level of discrete segments (i.e. 16x36) was chosen as it matches the sampling grid used in generating polar maps in FlowQuant®, a PET imaging software tool, and divides well into the 17-segment heart model

[0011] , It can be appreciated that other numbers of segments may be utilized.

[0100] The steps for acquiring these discrete myocardial segments are described below:

[0101] 1 . Using the LV long-axis, define a set of unit vectors aligned to the cardiac axes.

[0102] i. n1 , the unit vector along the LV long axis;

[0103] ii. n2, the unit vector perpendicular to the LV long axis pointing superior

[0104] iii. n3, the unit vector perpendicular to the LV long axis pointing lateral.

[0105] 2. For each desired (z,0) segment, where z is the nth layer from the basal plane of the ventricle (zero indexed), and 0 is the degrees clockwise around the LV long- axis from n2, create a wedge-shaped segmentation (as shown in FIG. 9). The distance from the apex to the center of the mitral valve is used to scale coordinates in the short-axis plane to ensure the wedge extends beyond the myocardium wall. The coordinates of the six vertices (in cartesian coordinates) are shown in FIG. 11 and defined by the following equations: ere I is the distance from the mitral valve center to the apex

[0106] 3. Find the intersection of the wedge segmentation with the myocardium segmentation (as shown in FIG. 10) using, for example the Slicer Segment Editor™, and label it according to its (z,0) position.3D Image Visualization3D PET Visualization

[0107] While the following example discusses PET visualization, it can be appreciated that the principles may extend to any imaging modality, that is, sampling within the discrete segments to identify a local value of that image type within the segment. This could be presented as a volume (coloring the myocardium segments) or on a surface object as is done with the ECGi.

[0108] In one example, an objective may be to create a single 3D visualization which shows a fusion image of both PET and ECGi. Since both imaging modalities are routinely encoded with color, fusion imaging poses a challenge: fusing color maps of two different imaging modalities is likely to result in misunderstanding or misperception due to the blending of two color maps. Instead, a segmentation was generated from the PET dataset based on intensity thresholding, which shows regions of the myocardium that met the threshold value and leave a void for regions that did not. This segmentation was viewed in fusion with the ECGi as a semitransparent white segmentation such that the color map of the ECGi could be perceived without significant distortion.

[0109] While this visualization is considered appropriate for defining regions of scar relative to ECGi in some cases, the visualization can be reliant on the chosen thresholds and septal regions of the PET were typically obscured by the ECGi. A semi-transparent ECGi object was also considered, but it was found that multiple semi-transparent objects in one view were difficult to understand and prone to misinterpretation. Additionally, the thresholded PET object loses valuable information in the gradient of uptake between scar tissue and healthy tissue, regions which are often the most prone to VT

[0012] ,

[0110] To resolve this, it was decided that two separate perspective- synchronized 3D views may be used, one displaying ECGi and the other displayed the PET object. Because the PET was not to be viewed in fusion with the ECGi, one can once again leverage color to encode PET values rather than using a single threshold. It was recognized that using the discrete segmentations of the myocardium from the CT, the PET intensity values were sampled within each discretesegmentation. It may be noted that this observation may extend to sampling of other modalities, such as EAM sampling within the same range (projection of surface onto CT segmentation surface). The maximum PET uptake value within each segmentation as a percentage of the global maximum uptake may be used to assign a color value to each segment. Colors are then assigned based on this normalized value using a rainbow color map.

[0111] When two PET types are loaded (FDG and Perfusion), the PET sampling is performed on both PET volumes, but only one PET type is used to color the volumetric PET object. By default, this is the FDG PET when available, as it is preferred for identifying regions of myocardial scar

[0013] , A dropdown menu can be provided to allow the discrete myocardial segments to be recolored based on the PET type that the user selects.Polar Map Visualization

[0112] A common visualization used for viewing nuclear imaging is a polar map. Polar maps are used to demonstrate the characteristics of the entire LV structure in a single 2D snapshot. Polar maps are found to have widespread use in nuclear cardiology and echocardiography.

[0113] Polar maps show apical information at the center, and basal information around the outside of the circle (similar to a polar projection of one hemisphere of the earth). By convention, the interventricular septum is centered at 9 o’clock, with the anterior and posterior walls shown at 12 and 6 o’clock, respectively.

[0114] In nuclear cardiology, the polar map is often used to present information about the radiotracer uptake in the left ventricle. The radiotracer uptake sampling can be exported from a sampling module (e.g., in the image registration and display module 46) in a 16x36 matrix corresponding to the (r,G) position within the polar map, where r is the ring number and theta denotes the rotational position in 10- degree increments clockwise from 0 at the top (anterior). To minimize or avoid differences in sampling between the sampling program and the discrete myocardium segmentation that could result in inaccuracies, in one example, a PET sampling process has been created based on the discrete myocardium segmentations. The values of each discrete segment are used to color the polar maps. The same rainbow color map can be used for both the polar map and volumetric PET object, providing a more intuitive link between visualizations.

[0115] These principles may extend to polar maps of EAM. Here, the system includes a projection from a 3D surface to polar maps, or samples in 3D as above (e.g., take point of largest voltage magnitude, earliest activation, etc., within the segment “wedge” in FIG. 9), or projects points from the EAM surface, sampled or otherwise, to the projection space as shown in FIG. 4.

[0116] Similarly, these principles may extend to providing polar maps of anatomic imaging (e.g. CT, MR, Echocardiography). This could be from sampling late imaging (LGE MR or LIE CT) within segments, or from looking at segment changes across cardiac cycle to present characteristics such as myocardial wall thickness, wall thickening, wall motion, etc. (which can be indicative of regions of scar or poor myocardial health).

[0117] The polar map can be divided into segments corresponding to the 17- segment heart model. The 17-segment heart model is a standardized nomenclature created by the AHA to describe discrete regions of the left ventricular myocardium, and is used in the current therapy planning process to communicate targets between imaging modalities. A bold line can be provided dividing each of the segments as shown in FIGS. 12A and 12B, and provide a hotkey (“f’) to toggle on / off the number associated with each segment. This is intended to promote knowledge translation and shared decision-making between clinicians during the planning process, especially for radiation therapy clinicians who are not as familiar with the heart anatomy and are unlikely to know the segment numbers of the 17- segment model.

[0118] The principles previously described for polar maps may be used to present information from a plurality of diagnostics, such as an overlay of two imaging modalities by means of image fusion such as alpha blending or similar.

[0119] The principles previously described for polar maps may extend to other projection views corresponding to other regions of the cardiac anatomy (e.g. atria, right ventricle, etc.).

[0120] It may be further noted that if the scan includes a time component (e.g., ECGi with 2000 timepoints), the visualization tool 12 can play the scan back temporally in the polar map views.Crosshair Position

[0121] A second hotkey can be provided (see “x” in FIG. 12A) to update the position of the crosshair in the polar map view. In this example implementation, acrosshair can be set in the 3D or 2D scene by pressing the Shift key and moving the cursor. The crosshair follows the cursor and is placed at the point of the cursor when the Shift key is depressed and changes the 2D slice view origin to the point of this crosshair. Using the hotkey, the polar coordinate representation of the crosshair position can be identified in the polar map view by a white ‘x’ symbol as seen in FIG. 12A. This was developed to further promote understanding of the relationship between the 3D and polar map visualizations. For example, if the user identifies a region of interest on EAM in 3D (see portion D in FIG. 4), they can place the crosshair at that position, and find the corresponding position in the polar map.

[0122] By nature, polar maps remove the transmural dimension of the myocardium. Similarly, when converting the 3D crosshair position to a polar map, the radial distance (i.e., the distance of the crosshair from the LV long axis) is lost.2D Slice Views

[0123] 2D slice views are widely used in reviewing medical imaging scans. PET is routinely reviewed in cardiac long-axis and short-axis slices to assess for cardiac defects such as reduced blood perfusion, and CT is most commonly evaluated using three orthogonal slices. 3D Slicer provides this functionality by default, opening a medical image volume in three orthogonal slice views. Within the proposed workflow, the familiarity of slice views can be leveraged for the assessment of CT and PET imaging, in addition to the polar map and 3D visualizations described earlier.Fusion Visualization

[0124] In the 2D slice views, it may be desirable to see a combination of multiple imaging scans (e.g., CT and PET) to provide information on both the high-resolution anatomy and the myocardial health simultaneously. At different steps in the proposed workflow, this can be achieved in different ways. During 3D and polar map target selection, the sampled PET discrete myocardium segments can be displayed in the 2D slice views as shown in FIG. 13A. The discrete myocardium segments are displayed with no outer contour and an opacity of 0.5, overtop of the cardiac CT slice view. If segments are selected as the ablation target, they will appear as an opaque black in the slice view. A toggle can be provided to turn on and off the visibility of the discrete myocardium segments, should the user prefer to interrogate the CT alone.

[0125] The cardiac CT may be displayed as the background image in greyscale, and the PET volume as the foreground image using a PET rainbow colourmap, though the user could select other colourmaps as desired. This is shown in FIG. 13B.

[0126] In either of the 2D slice visualizations, if the cardiac CT is a temporal CT (e.g., gated to the full cardiac cycle), it can be played back. This functionality may extend to any gated image where a temporal dimension is included (e.g., gated MR, gated PET).Cardiac Slice Views

[0127] In the slicer tool used, the images may be sliced into axial, sagittal, and coronal slice-views, relative to the patient body position. These planes, which may be referred to as “anatomic planes,” are commonly used in radiation therapy and medical diagnostics. When reviewing cardiac imaging, anatomic planes may be preferred to reorient the slice views to the position of the heart; that is to orient the three slice views to two orthogonal planes whose intersection is the LV long axis, and a third slice view orthogonal to the LV long axis, which may be referred to as “cardiac planes”.

[0128] The visualization tool 12 may provide a mechanism to toggle between anatomic planes and cardiac planes. This leverages the placement of the points at the apex and mitral valve center to define the LV long axis, and uses the normal vectors described above to define three planes. These planes are then used to resample the image volume(s) to provide slice views in the cardiac planes. Providing this feature and the ability to easily toggle back and forth between anatomic and cardiac planes not only provides the clinician with their preferred view for interpretation but is also intended to promote knowledge translation and shared decision-making between cardiology and radiation therapy clinicians as they can rapidly toggle back and forth.Target Selection

[0129] With the numerous visualizations offering a more thorough understanding of the relationship between the information provided by each individual imaging modality, the visualization tool 12 was configured to provide interactions with these visualizations that promote rapid and accurate target selection. This capability should include enabling selection of an ablation target, promoting shared decision making between clinicians, and using familiar interactions to promote rapid clinical adoption and user confidence.Polar Map Target Selection

[0130] Polar maps are a suitable intermediary for healthcare providers with different expertise, as they maintain a uniform shape and do not require interpretationof any patient-specific anatomical features. In the current therapy planning workflow, polar maps may be considered for single-modality characteristics, or any plurality of images by means of image fusion such as alpha blending or similar. The 17-segment model is used as a common frame of reference for clinicians to discuss their targets. Advantageously, the visualization tool 12 includes a unique ability to provide an interactive polar map used to facilitate target selection for RA. The correlation between the polar map view and the PET segmentation object allowed a simple relationship between selecting a target in the polar map view and updating the 3D and 2D Slice views accordingly. As noted above, and shown in FIGS. 14 and 15, polar maps could include an image fusion of two or more imaging modalities.

[0131] A method of target selection can be enabled which uses a mouse left click or left click+drag, adding all segments selected during the mouse click to the target. Effectively, this offers a “spray-paint” targeting method in the polar map view. The selected polar map cells are updated to black to indicate their addition to the target immediately after hovering the mouse over the cell with the left mouse button is depressed. Once the mouse click is released, the cells added to the target are then updated in the 3D and 2D slice views. This is achieved through the cell ID which correlates the cell IDs (R,O) from the polar map to the cell ID (z,0) of the discrete myocardium segments, adding the corresponding segment to the selected ablation target volume.

[0132] A progress dialogue box may then appear during the time required to calculate and recolor the PET object. A running list of target cells can be maintained to optimize the selection, only requiring cells not previously in the target to be added to the target. Similarly, cells can be removed from the target using an input (e.g., Ctrl+ left click or Ctrl+ left click+drag), with the same functionality as above. A list of the PET colors can be maintained so that when a cell is removed from the target, its color is changed from black back to the color corresponding to the PET intensity value.

[0133] To avoid processing delays, if present in the computing device used, the visualization tool 12 may be configured to show in the polar map which segments have been selected, but delay the update of the 3D scene until mouse release, when a dialogue box appears to denote scene update progress. A hotkey (e.g., spacebar) may be provided to toggle on / off the visibility of the target in the polar map view. This allows for a simple check to verify the underlying PET values of the selected target.

[0134] The ability to select a target in the polar map view and have it update a resulting target in the 3D scene addresses the objectives described earlier. Moreover, the use of polar maps as a standardized, familiar visualization for clinicians supports these objectives.

[0135] As discussed herein, the visualization tool 12 uniquely allows target selection in the polar map. This requires the image preparation to be performed in a certain manner (as described herein), and a way to “add back” a 3rddimension to translate from the polar map to 3D scene based on a user’s selection.

[0136] This feature is particularly advantageous since the color maps used for showing these data (e.g. PET, EAM, etc) may be difficult to overlay with diagnostic values. It is often difficult to correlate across multiple 3D views, but polar maps are standardized and can easily show the correlation between modalities. Allowing interaction in these maps allows a target to be selected far more simply and consistently than in 3D views (at least for some users).

[0137] In a first step, the PET-to-CT sampling correlation is created. Here, contrast enhanced cardiac CT (CECT) is registered to the corresponding image scan (PET, EAM, etc.). A volumetric model of the LV myocardium is extracted from CECT. For volumetric data (e.g. PET), the volumetric myocardium model is divided into a plurality of discrete volumes based on the orientation and position of the heart, identified by anatomical landmarks. For surface data (e.g. EAM), the heart surface is divided into a plurality of discrete surfaces based on the orientation and position of the heart, identified by anatomical landmarks. The imaging information from the non- CT image (e.g. PET, EAM) is sampled within each discrete segmentation volume or surface.

[0138] In a second step, the polar map is colored based on the values sampled in the first step above, i.e., with a mapping from (z, 0) in 3D to (r, 0) in 2D.

[0139] In a third step, the visualization tool 12 responds to user selection. Here, points or regions (e.g. cells, or click+drag to define a contour) are selected on a polar map are mapped back to the volumetric scene. This may be done by correlating an ID between each polar map segment to the corresponding volumetric segment. This could, however, be any method of mapping back to 3D such as saving an inverse coordinate transform from the polar map generation, to allow the selected point or region in the polar map scene to correspond to the volumetric scene.

[0140] Each point selected in a polar map corresponds to a point on the epicardial surface in 3D space, and the selection of multiple points allows the creation of a 3D contour. This 3D contour is then projected onto the nearest point of the outer surface of the myocardium segmentation. A volumetric target of the full width of the myocardium segmentation achieved in step 1 is segmented by taking the resulting 3D contour, and projecting this contour through the full myocardium thickness. It can be appreciated that the method described here is only one example and may be done on any ray originating at the centroid of the myocardium segmentation, and could equally be a ray perpendicular to the heart surface at each point on the contour, or similar.3D Target Selection

[0141] In an example workflow for 3D target selection, the following steps may be performed:

[0142] a) Point cloud placed by user;

[0143] b) Identify the point cloud centroid;

[0144] c) Project the point cloud on plane perpendicular to the line created between the point cloud centroid and the midpoint of the LV long axis (drawn between the mitral valve center and the LV apex);

[0145] d) Create 2D convex hull of the projected point cloud;

[0146] e) Map corresponding outer points back to 3D (by point index);

[0147] f) Create a spline connecting the set of 3D points that define the outer contour;

[0148] g) Sample the spline to determine the segment indices selected along the outer contour:

[0149] i. Convert to cylindrical coordinates;

[0150] ii. Z->R, theta -> theta;

[0151] Hi. Round theta to nearest 10 degrees, R to nearest 1 / 16th of the total distance from mitral valve center and the LV apex;

[0152] iv. Select each corresponding cell index in the polar maps;

[0153] h) Select all polar segments falling within the outer contour polar map representation.

[0154] It may be noted that as discussed herein, unlike typical methods of sampling for a polar map, the visualization tool 12 can use a mapping from cylindrical coordinates (R, theta, z) to polar (R, theta) where theta -> theta, z -> R, and the R dimension in cylindrical coordinates is ignored. It can be appreciated that typical polar map sampling methods could also be performed, such as bottlebrush sampling, which uses the present method for the basal % of the LV, then spherical sampling for the apical segments.

[0155] While polar maps may be the most appropriate visualization for standardized communication between clinicians with differing expertise, electrophysiologists may be more familiar with 3D rendered views in the context of VT ablation therapy. Through a catheter RF ablation system such as CARTO® (Biosense Webster®), ablation is performed entirely through 3D catheter position relative to a 3D invasive EAM and ablation points are displayed in this 3D view. A limitation identified with current solutions was the lack of an ability to accurately translate the regions of interest from EAM to other imaging modalities (or other diagnostics 18) used in therapy planning. By providing a 3D target selection tool in the target selection module 50, the visualization tool 12 can allow the electrophysiologist to identify a target based on the familiar EAM, which can then be interrogated relative to other imaging modalities by the imaging cardiologist.

[0156] One may constrain the selected volume to within the myocardium segmentation. This is functional and provides a reasonable correspondence between target selection on ECGi and the same region on CT but requires a thick spherical paintbrush to enable identification of the full myocardium thickness. Smaller paintbrushes in 3D would then require further refinement in 2D to define a fullthickness target. Additionally, painting the target over the ECGi obscured or distorted the ECGi map, depending on the color and transparency of the selected target region.

[0157] Often it is desirable to place a target and observe the temporal ECGi data in the targeted region, which may have been compromised by the placement of this target.

[0158] While this 3D target selection interaction may be configured to allow target selection on the 3D ECGi object, the functionality is equivalent in selecting the target on the surface of the 3D PET object. The ability to select a target in the 3D views and have it update display of the same target in the polar maps furthersupports the above objectives, and the use of 3D views familiar to the electrophysiologist for identifying a target region and correlating this target to 2D slice views and polar map views familiar to the imaging cardiologist also addresses the objectives.Right Ventricular Epicardium vs. Ventricular Septum Targeting

[0159] One limitation of ECGi is in the identification of arrhythmias with a septal origin because the ECGi maps represent the epicardial surface voltage and thus have limited information about septal electrical activity. Many clinicians acknowledge this limitation and exclude septal segments when comparing ECGi to invasive EAM or anatomical scar mapping [14-16], However, there is still meaningful information that can be gleaned from ECGi relating to septal activity, where Graham et al. demonstrated that VTs of septal origin could be measured with similar accuracy to those with origins elsewhere in the heart using ECGi

[0015] , In the workflow enabled by the visualization tool 12, if the target is painted to select regions on the right ventricular epicardial surface, this results in a selection of the ventricular septal myocardium at the same (R,0) position. This may deviate from the desired familiarity design requirement as there is no convention for correlating RV epicardial surface electrical information to ventricular septal myocardium electrical information, but it has been found that this was deemed the most intuitive response to a target painted on the right ventricular epicardium.

[0160] The workflow may also be based on the definition of a target within the left ventricle. VT originating from the right ventricular free wall is uncommon [17, 18] and thus was not a priority for compatibility in our workflow. The visualizations may still facilitate the understanding of a VT originating in the RV free wall but if an RV target is desired, it would need to be manually painted on the 2D or 3D views.Target Merging

[0161] Merging / adding targets may occur in real-time with suitable processing capabilities, or can be performed in a post-processing step for later integration into therapy planning tools such as radiation beam planning or catheter ablation systems. In the present workflow, both the 3D and polar map target selection tools effectively build a list of (R,0) segment IDs to include within the target. These segments can all be recolored to black replacing their PET color, but a new segmentation corresponding to the collection of these segments has not yet been created. Once the user is satisfied with the surface area extent of the target selected(not the transmural extent), the target can be merged into a single segmentation. Each cell ID within the target list is programmatically added to a new target segmentation using the 3D Slicer Segment Editor “Add” method. A resulting single segmentation object is generated which covers the full extent of the selected target.2D Slice-View Target Selection

[0162] Following the target selection and refinement in the polar map and 3D views, and the target merging into a single segmentation object, the user may choose to interrogate the target in the 2D slice-views. In this view, the clinician may determine that the automatic segmentation of the ventricular myocardium was insufficient and may add or remove from the target segmentation to adjust the transmural extent of the target. This is achieved using the 3D Slicer Segment Editor module paint and erase tools (with the spherical brush enabled). Ideally, these adjustments are unnecessary if the CT myocardium segmentation is sufficiently accurate but providing this backup adjustment functionality (and associated quality assurance check) is advantageous without a robust validation of an automated myocardium segmentation method. Adjusting the surface area extent of the target is also possible in 2D slice views and relies on slice-by-slice painting methods similar to those performed using current therapy planning software.Target Export to Radiotherapy Planning Software

[0163] Once the user is completely satisfied with the selected target, the target can be exported to DICOM-RT format for subsequent importing to radiation therapy treatment planning software (RTTPS) using the export function 52. This can be achieved using the 3D Slicer SlicerRT extension

[0019] , The target segmentation and Cardiac CT can be exported to a file format compatible with more RTTPS, where image registration can be performed to align the Cardiac CT and corresponding target to the radiation planning CT.Image Preparation and Targeting Workflow

[0164] Observations obtained from prototype demonstrations and interactions with clinicians led to the development of an associated workflow that may be used with the visualization tool 12. The tool 12 and its features were developed to allow iterative target selection and the ability to work back and forth between different features, notwithstanding the recommended workflow detailing the target selection process and the flow from coarse- to fine-target refinement. The specific steps and interactions are detailed below beginning with a high-level flow of the preparation ofthe visualizations prior to clinician use, as well as a recommended flow of the physician targeting process. The following example provides a workflow using PET / CT / ECGi, however, other combinations of diagnostics are possible, following a similar preparation workflow.Image Preparation Workflow

[0165] This section summarizes the steps used to prepare the visualizations for target selection. FIG. 16 illustrates the flow of image processing and registration to achieve the resulting image visualizations.

[0166] Load DICOM Data - the first step is to import all relevant DICOM format images into 3D Slicer®. The CECT may be imported using the 3D Slicer® MultiVolumeimporter extension to allow the temporal CT to be loaded as a volume sequence which can be played in a cine-loop.

[0167] Identify a Late Diastolic Phase for Targeting - play through the cardiacgated CECT to identify a suitable late-diastolic phase with limited imaging artefacts and a distended LV.

[0168] Crop CT Volumes - crop all CT volumes using the 3D Slicer® Crop Volume module to approximately the same rectangular extent as the contrast- enhanced cardiac CT.

[0169] Load ECGi Data - using the visualization tool’s ECGi import tool, import ECGi maps for the desired patient ID and VT cycle.

[0170] Define Anatomic Landmarks - place a marker at the apex and at the center of the mitral valve on the CECT.

[0171] Image Registration - using the SlicerElastix® toolbox, perform image registration. Use the CECT as the fixed image volume and perform a separate registration for each other CT. Set the output to a new transform, and apply that transform to the appropriate image (e.g. register ECGi CT to CECT, then apply that transform to the ECGi maps).

[0172] It may be noted that the workflow shown in FIG. 16A can leverage CT-to- CT image registration, wherein ECGi is acquired with a registered CT scan and PET scans are acquired with corresponding attenuation correction CET scans. The visualization tool 12 may leverage each of these CT scans to perform CT-to-CT registration, and then apply the resulting registration transform to the corresponding non-CT images.

[0173] For example, the routine for CT-to-CT registration may begin with a nondiagnostic CT being acquired with non-CT image types used in therapy planning of Radioablation. Here, CT / PET fusion image is acquired, where the CT is acquired for the purpose of attenuation correction. Non-contrast non-dynamic full-torso CT is acquired with the ECGI. For existing EAM modalities, this is only relevant for ECGi, since invasive mapping doesn’t have a prescreen CT, and surface-based registration methods would be used (i.e. not CT-to-CT).

[0174] Next, consider that these CT s are usually only used for the creation of the corresponding images. Instead, the visualization tool 12 can load them into a planning session. The system may register these CTs to the cardiac CT (at end- diastolic phase) using the Elastix® deformable image registration. The output of this registration is a transformation matrix (mapping of points from the original point space to a new point space (the cardiac CT point space).

[0175] The transformation matrix can be applied to the PET volume or to each EAM map to transform the non-CT images to the cardiac CT.

[0176] Myocardium Segmentation - create a new segmentation of a diastolic phase of the CECT. Define four segments (LV blood pool, myocardium, RV, and other anatomy) and define seed points for each segment. Use the 3D Slicer® Segment Editor Grow From Seeds tool. Adjust as necessary to achieve an accurate LV myocardium segmentation. Use a 2mm Gaussian filter in the 3D Slicer® Segment Editor tool to smooth the myocardium segmentation.

[0177] Discretize the Myocardium Segmentation - using the visualization tool 12, create the 576 discrete myocardium segments by selecting the “discretize myocardium” button.

[0178] Sample PET - for each PET type available (FDG and / or Perfusion), sample the PET within each discrete myocardium element using the 3D Slicer Segment Statistics module. Output the results as a table node for each PET type.

[0179] With respect to the segment statistics phase shown in FIG. 16A, it may be noted that polar map sampling based on CT segmentation registered to PET images provides a unique approach. It is found that most polar map sampling occurs based on perceiving a ventricular shape from the PET scan directly. The visualization tool 12 can use a CT scan segmentation, registered to the PET image, to define the sampling volumes used for generating the polar map details.

[0180] Prepare Visualizations - using the visualization tool 12, select the “Populate Polar Maps” button. This will load the ECG visualization for temporal ECGi playback, will color the discrete myocardium segments according to their PET values, and will generate PET polar maps. The target volume statistics shown in FIG. 16A shows the selected target volume and the percentage of total myocardium selected.

[0181] The 4D rendering provides a 4D ECGi object or an invasive EAM (or other non-invasive mapping technology) to recreate a familiar 3D reconstruction from EAM software. It may be noted, as discussed above, that while polar maps are widely used in nuclear imagine, the visualization tool 12 uniquely uses polar maps for ECGi, and moreover, provide temporal polar maps which allow playback of temporal ECGi data across the polar map.

[0182] The visualization tool 12 provides an enhanced platform to enable physician / clinician interaction, including, without limitation, the aforementioned target selection from polar maps, target selection by point-placement, and slice view conversion from cardiac orientation (long axis, short axis) to anatomic views (axi a l / sag itta l / co ro n a I) .

[0183] Referring to FIG. 16B, with respect to polar map targeting, selecting the target in the polar map updates the resulting target in the “volume” scene. Segments of the polar map are mapped back to a volumetric scene to select a volumetric target for ablation. Referring to FIG. 16C, with respect to 3D targeting, in this step, the clinician can place ablation points anywhere in the volumetric scene, which is later processed into an outer contour. This provides a unique approach to other methods which may require the outer contour to be selected. The selected volume may then be shown as a projection on the polar map as discussed earlier. Moreover, re-slicing the volumetric scene based on cardiac orientation provides a unique feature. It is recognized that many radiotherapy planning tools (and physicians) may only have standard anatomic views. Many cardiac clinicians are more familiar looking at cardiac axes. The visualization tool 12 allows the user to quickly switch between cardiac and standard anatomic slice views to promote knowledge translation between these clinicians with differing expertise.

[0184] This image preparation workflow has been developed for the specific set of images including ECGi, perfusion and FDG PET, and CECT. Several stages of the image preparation workflow have dependencies on previous stages of the workflow or on the availability of other imaging modalities. The discretization of themyocardium segment into smaller discrete segments as described above relies on both the placement of landmarks to define the LV long-axis and the segmentation of the full myocardium. The creation of the PET visualizations (polar maps and coloring of the discrete myocardium segments) relies on having discrete myocardium segments within which to sample the PET. Additionally, the mismatch PET polar map requires having both FDG and perfusion PET. This workflow requires performing these image preparation steps in this order and with these imaging modalities. However, the visualization tool 12 and associated workflow may be altered to enable other modalities used in differing orders.

[0185] FIG. 16D illustrates a variation on the workflow shown in FIG. 16A for EAM instead of ECGi. When compared to the ECGi workflow, the workflow shown in FIG. 16D does not have an associated CT with the electrical mapping to facilitate registration. As such, registration is performed using other techniques (e.g., manual, surface, fiducial, etc.). The diagram shown in FIG. 16D illustrates registration of the EAM to the CT myocardial segmentation. The other details are similar to FIGS. 16A- 16C described above and need not be repeated.

[0186] FIG. 16E illustrates a variation on the workflows shown in FIGS. 16A- 16D, where wall motion and wall thickness are calculated from a diagnostic. This may be beneficial for understanding potentially arrhythmogenic cardiac substrate, identifying an appropriate treatment region for radioablation of heart failure, and other applications. This figure demonstrates the use of echocardiography, but a similar workflow may be employed for any imaging modality that can identify the cardiac anatomy throughout the cardiac cycle, such as MRI, CT, etc. This process details only the measurement of wall thickness but this may be used in addition to any of the other aspects of image preparation demonstrated in FIGS. 16 A-D. Sampling of wall motion may occur within discrete segments as is done for PET sampling, averaging wall thickness, thickening, or motion across the cardiac cycle within each discrete myocardial segment defined by similar methods to those shown in FIGS. 9-11 .

[0187] Referring to FIGS. 23A and 23B, two illustrations are provided to show that wall motion and wall thickness calculations may be measured on ECHO / CT / MR, etc. The distance from the cardiac long-axis to the endocardium and epicardium are shown in green (200) and red (202), respectively. Myocardial wall thickness is a measurement from the left endocardial to epicardial surfaces (or left endocardial to right endocardial surfaces when measuring the interventricular septum). This may be measured perpendicular to the LV long-axis as shown in FIGS. 23A and 23B, orperpendicular to the endocardium, and is measured in one or several locations within each myocardial segment and is averaged. Myocardial thickening is a measurement of the change of myocardial thickness across the cardiac cycle (end-diastolic vs end- systolic or min thickness vs max thickness). Myocardial wall motion is measured as the change in distance to the left ventricular long-axis across the cardiac cycle (end- diastolic vs end-systolic or min thickness vs max thickness).

[0188] FIG. 24 shows a similar wall thickness, thickening, and motion measurement on a short-axis MRI slice. The resulting myocardial thickness, thickening, or wall motion measurements may be displayed in 4D view polar maps, and other visualizations as is done with PET values in previously described methods.

[0189] It can be appreciated that while the examples discussed herein focus on LV VT RA treatment, other applications are possible by applying the principles discussed. For example, RA of RV VT or PVCs, RA of left and right atria for atrial arrhythmias, RA of cardiac sarcoidosis, RA of myocardium / septum for hypertrophic obstructive cardiomyopathy (HOCM), planning for invasive catheter ablation (e.g., to shorten the procedure time by identifying the region for ablation and eliminating the need for catheter mapping), or other applications wherein there is a need for connecting electrical mapping to other cardiac imaging.

[0190] It can also be appreciated that further applications are possible by applying the principles discussed, such as in planning heart failure treatment. In such an application, electrical information may not be required, but rather anatomic information may be sufficient for planning the heart failure treatment.Target Selection Workflow

[0191] This section presents a recommended workflow through the visualizations and associated Ul, as shown in FIG. 4. The workflow recommends starting with high-level region selection based on 3D ECGi data, then refining the target according to PET information in polar maps, then making fine adjustments to the target informed by PET-CT fusion imaging in 2D Slice views. This is not a mandatory flow, for example, the clinician may choose to define the target initially in a polar map view; however, based on observations and discussions with our clinician group, this was perceived to be an effective flow of information for targeting. This is not intended to act as a user manual but rather as an overview of the flow of user interactions through the visualizations provided for target selection.

[0192] Select appropriate leads for ECG - default leads are selected to represent a 3- lead ECG. In most cases these are sufficient, but if any of the leads have poor signal quality, select alternative leads to recreate the 3-lead ECG.

[0193] Evaluate ECGi - evaluate the different ECGi map types, evaluating each in 3D. For the potentials and propagation maps, evaluate temporal data using the ECG playback tools. Determine the anticipated target region and visually compare this region to the 3D PET object in the synchronous view.

[0194] 3D Target Selection - place points sequentially around the outer contour of the desired target in 3D, either on the ECGi or PET 3D surface.

[0195] Adjust Target in Polar Maps - based on the target from the 3D target selection above, review the target relative to the polar map views and adjust the target accordingly.

[0196] Merge Target - once satisfied with the approximate surficial extent of the target, select the “Merge Target” button to stitch all target sub-segments into a single target segmentation.

[0197] Refine Target in 2D Slice-Views - compare the target to the CECT and PET fusion image in 2D slice views and adjust as needed.

[0198] Export Target to DICOM-RT - export the target and CECT to RTTPS.

[0199] The exported DICOM-RT format targets from this workflow, along with the CECT, are exported for import to RTTPS. This is intended to work as a precursor to any RTTPS beam planning. Cumulatively, the development of these visualizations and workflow created a complete package for defining an ablation target for RA. It can be appreciated that the radiation treatment planning CT or MRI could also be registered into the data scene, and the target could be exported on that image rather than the CECT. That is, the target could be exported onto any image volume.Illustrative Workflow

[0200] Referring now to FIGS. 17-22, an overview of an example workflow for utilizing the visualization tool 12 is shown. As shown in FIG. 17, to begin, the user may select control panel settings. Here, within the control panel, image selection is made, which updates the 2D and 3D scenes in the GUI. Then, the ECG slider may be used to update or select the temporal ECGi information. By selecting the “Select Target in 3D” button, the user may select the target area in the 3D scene. The target volume statistics are updates and the polar maps may be used to select the targetregion and display PET and ECGi polar maps. By selecting the “Merge Target” button, the individual target segments are stitched together once the user is satisfied with the polar map targeting.

[0201] Next, in FIG. 18, the ECG selector may be used to change the VT Cycle and EAM map type. The ECG Slider can be used to drag through temporal EAM data and the play button and skip frame can be used to control temporal EAM. The ECG lead selection can be adjusted (e.g., see FIG. 6 for a map of lead IDs).

[0202] Referring now to FIG. 19, the user may click and drag to rotate the 3D scene (which may also be done during the EAM playback). In this example, the user may hold the “shift” key while hovering the cursor over the 3D scene to move the crosshair and 2D slice view origin to that point. This may be done to correlate EAM locations to regions of interest on PET and CT images.

[0203] In a second stage, shown in FIG. 20, once EAM is evaluated sufficiently to target an approximate region of interest in 3D, the user may press a “select target in 3D view” button to begin selecting a region in the 3D view; then click to place markers on the 3D scene to define the boundary of the target; double click to complete the closed curve, and press “done”. The boundary of the target then appears on the 3D PET object. It may be noted that targets painted on the epicardial surface of the right ventricle may be projected to the ventricular septum, projecting radially in towards the LV long-axis.

[0204] A third stage is shown in FIG. 21. Here, it is illustrated that the user may scroll down in the control panel as needed to view polar maps. If satisfied with approximate contour defined in the third stage (described above), the user may click and drag on the polar maps within the black contour to fill in the target. Otherwise, the user may click “undo” to detect the 3D curve and repeat the third stage. FIG. 21 also shows that considering values of PET uptake and the resultant target size can lead to adding or removing segments to the target using the polar maps. For example, “click” or “click+drag” may be used to add while “ctrl+click” or “ctrl+click+drag” may be used to remove. Various hotkeys may be implemented, for example, using the spacebar to toggle on / off target visibility in polar map views, “x” to update the crosshair in the polar map views to correspond to the crosshair in the 3D scene, and “f” to toggle on / off the IDs of the 17-segment heart model for reference.

[0205] A stage not shown in the figures, allows the user to merge the target. For example, once satisfied with the polar map targets, the user may select the “mergetarget” button (e.g., scroll down below the polar maps in the control panel) to merge all individual selected segments into one target volume. Then, the user may make the following changes for 2D target refinement: turn off PET in 2D, turn on alpha blending of PET with CECT, switch to four-up view, copy target from 16x36 ventricle segmentation to target segmentation, adjust ECGi transparency as desired to see corresponding volume, and switch to segment editor.

[0206] In a fourth stage, shown in FIG. 22 target refinement may be performed in 2D, e.g., to add and remove to the target in 2D views until satisfied with the result. A brush diameter may also be adjusted as highlighted in FIG. 22.Computing and User Interaction Efficiencies

[0207] One of the major limitations of the current therapy planning process is the time required by multiple clinicians to determine an ablation target. Anecdotally, it has been found that it can take between 1-3 hours for each case, which does not include individual clinician time preparing content and reviewing images for these meetings. To demonstrate the efficiencies created by executing and leveraging the visualization tool 12 described herein, the timing of the target selection workflow has been measured, including specific steps in the workflow, to measure the potential time-saving benefits of such a tool and to identify aspects of the therapy planning that could be further improved. The timing results are presented below in Table 4.Table 4: Timing of Steps within the Targeting Workflow

[0208] These results demonstrate that this software tool can be a timesaving solution for clinicians. Even in their first exposure to this software tool, they were able to perform therapy planning faster than the current methods.

[0209] The results in Table 4 are extracted directly from the start and end points of specific steps. These include the time required for discussion and training during the case planning sessions. Often, the clinicians would pause to explain their decision-making or a clinical concept prior to proceeding with the therapy planning. If one removes these pauses, the average case planning times were 21.5 minutes for a user’s first case and 17.9 minutes for a user’s second case.Heart Failure Patient DataFIG. 25 shows the ejection fraction change in patients that received RA for treatment of VT. This example data supports the notion that RA could be beneficial for improving ejection fraction and possible RA use in heart failure.Case Planning Times

[0210] Referring to FIG. 26, a study was performed using the presently described software tool where a number of electrophysiologists with no prior experience in RA planning performed targeting on a number of retrospective cases. After only a single training session, the users operated the tool independently and were able to achieve comparable accuracy to expert users not having access to the software tool, in less time. The study is summarized in greater detail below.

[0211] Ventricular arrhythmias are the leading cause of sudden cardiac death, accounting for the majority of 60,000 cardiac arrests annually in Canada. Stereotactic arrhythmia radiotherapy (STAR) is an emerging and completely non-invasive treatment option for ventricular arrhythmia ablation. Early clinical results show promise, but existing treatment planning solutions are lacking, resulting in delayed treatment and high interobserver targeting variability.

[0212] To evaluate the accuracy, precision, and duration of STAR targeting using the tool 12, a dedicated precision targeting multi-modality imaging tool.

[0213] Five electrophysiologists (EP) with no prior STAR experience, used the tool 12 to perform repeat targeting in 10 previously treated patients (6 ICM, 4 NICM, VT-free @1-year post-STAR). Imaging included contrast-enhanced CT, perfusion and viability PET, and electrocardiographic imaging (ECGi), all registered into the same frame-of-reference usingthe tool 12. One case was randomly selected for additional intra-operator repeatability assessment. Total planning time was measured from initial case-loading into the tool 12 until target-exporting to RT-DICOM format. Average Hausdorff Distance (HD) and Dice Similarity Coefficient (DSC) were measured for inter- and intra-user variability, as well as target volume for accuracy compared to the clinical standard targets used for treatment. Each EP completed a NASA-TLX survey at the study completion to assess the treatment planning task load in six categories (mental, physical, and temporal demand, performance, effort, and frustration).

[0214] Average case planning time was 11 min ± 7 min. Task load was reduced compared to reported values for baseline planning methods in all six categories (p<0.05). Novice users’ target volumes tended to be smaller than the clinical standard volumes (58 ± 30 cc vs 73 ± 30 cc, p=0.11). The average intra-user HD was 4 ± 2mm with DSC of 0.72 ± 0.13, whereas inter-user HD was much higher at 14 ± 9 mm, with lower DSC of 0.41 ±0.16.

[0215] The tool 12 was considered easy to use and facilitated rapid target acquisition by electrophysiologists without STAR experience. Novice users tended to select smaller treatment volumes with inter-user variability using the tool 12 comparable to prior reports from expert users without a dedicated tool. The tool 12 may expedite physician training and treatment planning with comparable accuracy vs current standard methods.

[0216] It can be appreciated that beyond radioablation, the visualization tool 12 may be adapted for other applications, e.g., for planning catheter ablation therapies in advance of catheter ablations. This could enable the process to occur faster, potentially close to realtime, so as to allow an invasive map to be acquired, integrated with other pre-acquired image types, then analyzed on-site prior to the administration of the ablation. This recognizes that minutes saved in planning and accessible interpretations of images can be critical, as the patient is on the table under anesthetic during this time.

[0217] For simplicity and clarity of illustration, where considered appropriate, reference numerals may be repeated among the figures to indicate corresponding or analogous elements. In addition, numerous specific details are set forth in order to provide a thorough understanding of the examples described herein. However, it will be understood by those of ordinary skill in the art that the examples described herein may be practiced without these specific details. In other instances, well-known methods, procedures and components have not been described in detail so as not to obscure the examples described herein. Also, the description is not to be considered as limiting the scope of the examples described herein.

[0218] It will be appreciated that the examples and corresponding diagrams used herein are for illustrative purposes only. Different configurations and terminology can be used without departing from the principles expressed herein. For instance, components and modules can be added, deleted, modified, or arranged with differing connections without departing from these principles.

[0219] It will also be appreciated that any module or component exemplified herein that executes instructions may include or otherwise have access to computer readable media such as transitory or non-transitory storage media, computer storage media, or data storage devices (removable and / or non-removable) such as, for example, magnetic disks, optical disks, or tape. Computer storage media may include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information, such as computer readable instructions, data structures, program modules, or other data. Examples of computer storage media include RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transitory computer readable medium which can be used to store the desired information and which can be accessed by an application, module, or both. Any such computer storage media may be part of the computing environment 10, any component of or related thereto, etc., or accessible or connectable thereto. Any application or module herein described may be implemented using computer readable / executable instructions that may be stored or otherwise held by such computer readable media.

[0220] The steps or operations in the flow charts and diagrams described herein are provided by way of example. There may be many variations to these steps or operations without departing from the principles discussed above. For instance, the steps may be performed in a differing order, or steps may be added, deleted, or modified.

[0221] Although the above principles have been described with reference to certain specific examples, various modifications thereof will be apparent to those skilled in the art as having regard to the appended claims in view of the specification as a whole.References

[0222] [1] B. A. Koplan and W. G. Stevenson, "Ventricular tachycardia and sudden cardiac death," (in eng), Mayo Clinic proceedings, vol. 84, no. 3, pp. 289-297, 2009, doi: 10.1016 / S0025-6196( 11 )61149-X.

[0223] [2] J. A. Fallavollita, B. J. Malm, and J. M. Canty, "Hibernating MyocardiumRetains Metabolic and Contractile Reserve Despite Regional Reductions in Flow, Function, and Oxygen Consumption at Rest," Circulation research, vol. 92, no. 1 , pp. 48-55, 2003, doi: 10.1161 / 01. RES.0000049104.57549.03.

[0224] [3] S. M. Al-Khatib et al., "2017 AHA / ACC / HRS Guideline for Management ofPatients With Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death: A Report of the American College of Cardiology / American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society," Journal of the American College of Cardiology, vol. 72, no. 14, pp. e91-e220, 2018, doi: 10.1016 / j.jacc.2017.10.054.

[0225] [4] C. G. Robinson et al., "Phase l / ll Trial of Electrophysiology-GuidedNoninvasive Cardiac Radioablation for Ventricular Tachycardia," Circulation, vol. 139, no. 3, pp. 313-321 , 2019 / 01 / 15 2019, doi: 10.1161 / CIRCULATIONAHA.118.038261 .

[0226] [5] C. Wei et al., "Non-invasive stereotactic radioablation: a new option for the treatment of ventricular arrhythmias," (in eng), Arrhythmia & electrophysiology review, vol. 8, no. 4, pp. 285-293, 2020, doi: 10.15420 / aer.2019.04.

[0227] [6] C. Redpath et al., "Non-invasive cardiac radioablation for recurrent ventricular tachycardia: initial results and insights," Canadian journal of cardiology, vol. 37, no. 10, pp. S41-S41 , 2021 , doi: 10.1016 / j.cjca.2021 .07.087.

[0228] [7] H. J. Wellens, F. W. Bar, and K. I. Lie, "The value of the electrocardiogram in the differential diagnosis of a tachycardia with a widened QRS complex," (in eng), American Journal of Medicine, vol. 64, no. 1 , pp. 27-33, Jan 1978, doi: 10.1016 / 0002-9343(78)90176-6.

[0229] [8] D. Mattes et al., "PET-CT image registration in the chest using free-form deformations," IEEE Transations on Medical Imaging, vol. 22, no. 1 , pp. 120- 128, 2003, doi: 10.1109 / TM 1.2003.809072.

[0230] [9] A. Khalil et al., "An overview on image registration techniques for cardiac diagnosis and treatment," (in eng), Cardiology research and practice, vol. 2018, pp. 1437125-1437125, 2018, doi: 10.1155 / 2018 / 1437125.

[0231]

[0010] D. M. Zhang et al., "Cardiac radiotherapy induces electrical conduction reprogramming in the absence of transmural fibrosis," (in eng), Nature Communications, vol. 12, no. 1 , p. 5558, Sep 24 2021 , doi: 10.1038 / S41467-021-25730-0.

[0232]

[0011] R. Klein et al., "Fully automated software for polar-map registration and sampling from PET images," in 2006 IEEE Nuclear Science Symposium Conference Record, 29 Oct.-1 Nov. 2006 2006, vol. 6, pp. 3185-3188, doi: 10.1109 / NSSM IC.2006.353687.

[0233]

[0012] K. Soejima et al., "Electrically unexcitable scar mapping based on pacing threshold for identification of the reentry circuit isthmus: feasibility for guiding ventricular tachycardia ablation," (in eng), Circulation, vol. 106, no. 13, pp. 1678- 83, Sep 24 2002, doi: 10.1161 / 01 .cir.0000030187.39852. a7.

[0234]

[0013] H. Chalian et al., "Incremental value of PET and MRI in the evaluation of cardiovascular abnormalities," (in eng), Insights Imaging, vol. 7, no. 4, pp. 485- 503, Aug 2016, doi: 10.1007 / sl 3244-016-0494-5.

[0235]

[0014] P. S. Cuculich et al., "The electrophysiological cardiac ventricular substrate in patients after myocardial infarction: noninvasive characterization with electrocardiographic imaging," Journal of the American College of Cardiology, vol. 58, no. 18, pp. 1893-1902, 2011 , doi: 10.1016 / j.jacc.2O11.07.029.

[0236]

[0015] A. J. Graham et al., "Evaluation of ECG imaging to map hemodynamically stable and unstable ventricular arrhythmias," (in eng), Circulation: Arrhythmia and Electrophysiology, vol. 13, no. 2, p. e007377, Feb 2020, doi: 10.1161 / circep.119.007377.

[0237]

[0016] A. J. Graham et al., "Assessing noninvasive delineation of low-voltage zones using ECG imaging in patients with structural heart disease," (in eng), JACC Clinical Electrophysiology, vol. 8, no. 4, pp. 426-436, Apr 2022, doi: 10.1016 / j.jacep.2021 .11 .011 .

[0238]

[0017] S. M. D. Menon et al., "Right ventricular tachycardia," vol. 128, ed.Circulation: AHA Journals, 2013.

[0239]

[0018] W. Shimizu et al., "Ventricular tachycardia originating from the right ventricular free wall in a patient with an old myocardial infarction," Chest, vol. 100, no. 1 , pp. 276-277, 1991 / 07 / 01 / 1991 , doi: 10.1378 / chest.100.1 .276.

[0240]

[0019] C. Pinter et al. , "SlicerRT : radiation therapy research toolkit for 3D Slicer,"(in eng), Medical Physics, vol. 39, no. 10, pp. 6332-8, Oct 2012, doi: 10.1118 / 1.4754659.

[0241]

[0020] van der Ree MH, Cuculich PS, van Herk M, Hugo GD, Balt JC, Bates M, HoG, Pruvot E, Herrera-Siklody C, Hoeksema WF, Lee J, Lloyd MS, Kemme MJB, Sacher F, Tixier R, Verhoeff JJC, Balgobind BV, Robinson CG, Rasch CRN, Postema PG.Interobserver variability in target definition for stereotactic arrhythmia radioablation. Front Cardiovasc Med. 2023 Sep 20; 10: 1267800. doi: 10.3389 / fcvm.2023.1267800. PMID: 37799779; PMCID: PMC10547862.

[0242]

[0021] Hohmann S, Henkenberens C, Zormpas C, Christiansen H, Bauersachs J,Duncker D, Veltmann C. A novel open-source software-based high-precision workflow for target definition in cardiac radioablation. J Cardiovasc Electrophysiol. 2020 Oct;31(10):2689- 2695. doi: 10.1111 / jce.1466O. Epub 2020 Jul 21 . PMID: 32648343.

Claims

Claims:1 . A method of registering data of a first diagnostic to at least one other diagnostic, the method comprising: visualizing the first diagnostic relative to the at least one other diagnostic.

2. The method of claim 1 , wherein registration of a non-computed tomography (CT) diagnostic is obtained by registering a CT scan that is registered to the non-CT diagnostic and applying a resulting transformation to the non-CT diagnostic.

3. The method of claim 2, wherein the non-CT diagnostic is positron emission tomography (PET) or single photon emission computer tomography (SPECT) and the preregistered CT scan is an attenuation correction CT.

4. The method of claim 2, wherein the non-CT diagnostic is electroanatomic mapping(EAM), and the pre-registered CT scan is the CT scan used in generating EAM geometry.

5. The method of claim 1 , wherein the first diagnostic is EAM.

6. The method of claim 1 , wherein the at least one other diagnostic is CT.

7. A method comprising: obtaining a 2D projection view of diagnostic scans; correlating the projection view to one or more 3D scenes; and displaying correlated data in a 3D imaging view.

8. The method of claim 7, wherein the projection view is a polar map.

9. A method comprising: using one or more segmentations from an imaging scan to sample the scan or other registered imaging scans within each segment; and using the sampling to generate one or more visualizations.

10. The method of claim 9, wherein the one or more visualizations comprises 3D objects or polar maps.11 . The method of claim 9, comprising using a segmentation from a CT scan to sample the other registered imaging scan, wherein the segmentation from the CT scan has been separated into a plurality of discrete segments.

12. The method of claim 11 , wherein the other registered imaging scan comprises EAM, MRI, PET, or ultrasound.

13. A method comprising: selecting a volumetric target through a projection view by maintaining a correlation between a 3D scene to the projected view to maintain a dimension from the projection.

14. The method of claim 13, wherein the projection view is obtained using sampled information from one or more scan types.

15. The method of claim 14, wherein the one or more scan types comprise EAM, MRI, PET, or ultrasound.

16. The method of claim 13, wherein the volumetric target is delineated in the projection view and is correlated to a resulting transmural target by a preserved mapping.

17. A method comprising: selecting one or more targets in any one of a plurality of visualizations; and having the target update in other ones of the plurality of visualizations to show a corresponding selected region / volume.

18. The method of claim 17, wherein the target comprises an ablation target selected in one view and a corresponding target is displayed in the plurality of other visualizations.

19. The method of claim 18, wherein the ablation target is selected in a polar map and projected back to a 3D scene.

20. The method of any one of claims 17 to 19, wherein a selected target is motion- tracked throughout a cardiac cycle to identify a summative volume occupied by the selected target throughout the cardiac cycle.21 . The method of any one of claims 1 to 20, comprising providing an ability to assess in anatomic and cardiac slice views.

22. The method of any one of claims 1 to 21 , comprising providing an ability to export or integrate a selected target volume into a planning or treatment system.

23. The method of any one of claims 1 to 22, wherein two temporal diagnostics are registered temporally in a cardiac cycle.

24. A method comprising: segmenting a myocardium across a plurality of phases of a cardiac cycle to identify at least one of wall thickness, wall thickening, and wall motion, with a segmentation that has been separated into a plurality of discrete segments.

25. The method of claim 24, wherein the plurality of discrete segments is obtained using a segmentation from a CT scan to sample the other registered imaging scan, wherein the CT scan has been separated into the plurality of discrete segments.

26. A computer system comprising: a processor; and a memory, the memory storing processor executable instructions that, when executed by the processor, cause the computer system to perform the method of any one of claims 1 to 25.

27. A computer-readable medium storing processor executable instructions that, when executed by a processor of a computer system, cause the computer system to perform the method of any one of claims 1 to 25.

Images