Devices, methods, and computer programs for segmenting membrane septa

Abstract

The present invention relates to a device and method for segmenting a membrane septum. The method comprises: segmenting the left ventricular outflow tract (LVOT) of the heart in a 3D image dataset; determining wall thickness information indicating the wall thickness of the membrane septum at different locations in the segmented left ventricular outflow tract toward the right ventricle and right atrium; mapping the determined wall thickness information to the surface of the aforementioned portion of the segmented left ventricular outflow tract; and segmenting the membrane septum in the 3D image dataset based on the mapped wall thickness information.

Description

Technical Field

[0001] The present invention relates to a device, method, and computer program for segmenting a membranous septum.

Background Art

[0002] The so-called membranous septum (i.e., the membranous fibrous component of the entire cardiac septum) is a small structure within the heart located between the left ventricular outflow tract (LVOT) and the right ventricle / right atrium, and is characterized by a thin wall between the left and right ventricles at the height of the tricuspid valve annulus. It indicates the location of the conduction pathways within the heart. Certain geometric properties of the membranous septum are associated with the risk of requiring a pacemaker after transcatheter aortic valve replacement (TAVR).

Summary of the Invention

Problems to be Solved by the Invention

[0003] The membranous septum is a small, unobtrusive non-planar structure that is difficult to find and annotate in a 3D image dataset, such as a dataset of cardiac computed tomography (CT) angiography (CTA) images, particularly in a standard planar slice view. Furthermore, anatomical knowledge is available, which, when appropriately used, helps to identify the membranous septum. As such, large datasets (such as those required for neural network-based approaches) are avoided along with the tedious and hence costly manual annotation.

[0004] The paper "Virtual septal myectomy for preoparative planning in hyperhropic cardiomyopathy", published in The Journal of Thoracic and Cardiovascular Surgery, vol 158, no.2, 2019-08-01, pages 455-463 by Takayama Hiroo et al., discloses a virtual resection (VM) technique using three-dimensional reconstruction of gated cardiac CT, supporting the objective intraoperative evaluation of the validity of resection. [Means for solving the problem]

[0005] The object of the present invention is to provide a device, method, and computer program for efficiently and reliably segmenting membrane septa in an automated manner.

[0006] In a first aspect of the present invention, a device for segmenting a membrane septum is presented, which device In a 3D image dataset, the left ventricular outflow tract (LVOT) of the heart is segmented. The wall thickness information indicating the thickness of the membranous septum wall at different locations in the segmented LVOT toward the right ventricle and right atrium is determined, and The determined wall thickness information is mapped onto the surface of the segmented portion of the LVOT, and the membrane septum is segmented in the 3D image dataset based on the mapped wall thickness information. It has a circuit configured as follows.

[0007] Further aspects of the present invention provide a corresponding method, a computer program having program code means for causing a computer to perform steps of the method disclosed herein when executed on a computer, and a non-temporary computer-readable recording medium storing the computer program performing the method disclosed herein when executed by a processor.

[0008] Preferred embodiments of the present invention are defined in the dependent claims. The claimed methods, computer programs and media are understood to have preferred embodiments similar to and / or identical to the claimed systems, particularly those defined in the dependent claims and disclosed herein.

[0009] The present invention is based on the idea of ​​enabling automated segmentation of the membrane septum by segmenting the LVOT and using this segmented LVOT to determine the thickness information of the septum adjacent to the LVOT on the lateral sides of the right ventricle and right atrium. Based on this thickness information, the membrane septum with the smallest thickness can be identified, in particular, as part of the entire cardiac septum. Mapping the region of interest to a two-dimensional reformat can simplify subsequent image processing steps for manual and / or automated depiction and quantification of the membrane septum.

[0010] According to one embodiment, the circuit is further configured to segment the membrane septum by applying one of the following to the surface of the segmented portion of the LVOT: thresholding, independent component analysis, and k-neighborhood clustering. Generally, any standard technique for segmentation is applied to the surface of the LVOT for this purpose.

[0011] The circuit is further configured to segment the LVOT using model-based segmentation in order to fit the heart model to a 3D image dataset. This provides a good starting point for subsequent steps in the process.

[0012] In a preferred embodiment, the cardiac model has an index of the region where the membranous septum should be found, and / or information that enables the generation of two or more planes perpendicular to the LVOT and / or the determination of the aortic valve annular surface. This improves the segmentation of the membranous septum. Thus, the information may include geometric information in the form of landmarks having coordinates such as, for example, the centroid of the aortic valve, the normal vector of the aortic valve, the normal vector of the aortic valve annular surface, the normal vector of the ascending aorta, and / or the centroid of the septum.

[0013] In a practical implementation, the circuit described above is: Determining an axis that approximates the center line of the aforementioned LVOT, Defining a series of rays perpendicular to the axis and / or surface, passing through the wall of the LVOT in the segmented portion of the LVOT, To determine the corresponding wall thickness information for the aforementioned series of light rays, The wall thickness information is mapped to the position of the surface through which each of the aforementioned light rays passes the LVOT wall. This may be further configured to determine wall thickness information.

[0014] As a result, the circuit, Before determining the aforementioned wall thickness information, the surface of the aortic valve annularity is determined, Defining an axis perpendicular to the aortic valve annular surface and The system is further configured to determine the axis.

[0015] In another embodiment, the circuit is further configured to determine the axis by fitting a line passing through the centroids of two or more surface rings of the ascending aorta, in particular, as defined by a fitted geometric model (model-based segmentation).

[0016] Furthermore, in another embodiment, the circuit is further configured to determine, as wall thickness information, the wall thickness value of the septum along each ray, or an alternative value of this wall thickness value.

[0017] Thereby, as an alternative value of the wall thickness value, an average image value (or pixel value or gray value) of the ray segments of each of the rays starting from the surface of the LVOT and having a predetermined or adaptable length is obtained. This image value is related to radiopacity (radiation opacity, typically Hounsfield unit (HU)), which correlates (almost linearly) with the mass density (with respect to water and air). Thus, for a given assumed material, the accumulated density yields an approximate mass that is converted to length (thickness).

[0018] Furthermore, the wall thickness value is determined based on the length of the ray segments of each of the rays starting from the surface of the LVOT and ending with a change in the image value (or pixel value or gray value) indicating a change in tissue.

[0019] According to another embodiment, the wall thickness information is determined by using a trained algorithm or computer system, particularly a trained machine learning, trained classifier or trained neural network, trained to estimate the thickness of the wall of an organic structure in a 3D image dataset of the heart. This does not require the application of a dedicated algorithm but enables the use of artificial intelligence for this task.

[0020] The circuit may be further configured to determine an axis by calculating a distance map inside the LVOT and maximizing the straight-line distance from the surface of this LVOT within the LVOT.

[0021] According to another embodiment, the circuit segments the right ventricle and the right atrium, finds the point closest to the LVOT, identifies the ray passing through this point, defines an angular range for rotating the ray around the axis and a translational range for moving the ray up or down the axis, and is further configured to determine the region in which the membranous septum should be found.

[0022] According to yet another embodiment, the circuit utilizes a geometric model represented by a triangular mesh to identify, for each ray directed outward from the faces of the polyhedron of the LVOT, the 3D coordinates where this ray intersects one of the mesh triangles that intersects the vertices of the triangle, and uses the 3D coordinates as the starting point of the ray for estimating the local thickness of the septum and is further configured to perform the above.

[0023] The circuit may be further configured to visualize the determined wall thickness information and / or the segmented membranous septum on the surface of the segmented portion of the LVOT or on a flattened reformatted version of the segmented portion of the LVOT. This helps the user understand the results of the segmentation and make a diagnosis.

Brief Description of the Drawings

[0024] These and other aspects of the invention will become apparent from the embodiments described below and will be described with reference to these. [Figure 1] FIG. 1 shows a schematic view of a part of the anatomical structure of the human heart. [Figure 2] FIG. 2 shows a schematic view of an embodiment of the device according to the invention. [Figure 3] FIG. 3 shows a flowchart of an embodiment of the method according to the invention. [Figure 4] FIG. 4 shows a flowchart of another embodiment of the method according to the invention. [Figure 5] FIG. 5 shows a schematic view of the left ventricle with a heart model and additional information encoded. [Figure 6] FIG. 6 shows two cross-sectional views through the heart along different planes. [Figure 7] FIG. 7 shows a cross-sectional view showing the search space in the z direction. [Figure 8] FIG. 8 shows cross-sectional views from different viewpoints. [Figure 9]Figure 9 shows a cross-sectional view of the search space in the direction of rotation. [Figure 10] Figure 10 shows the results of an MBS-based implementation that gives a polyhedron / map covered as a texture. [Figure 11] Figure 11 shows a polyhedron / map using the occupied regions of the right ventricle and right atrium. [Figure 12] Figure 12 shows another exemplary embodiment of the present invention. [Modes for carrying out the invention]

[0025] Figure 1 shows a schematic diagram of the major parts of the anatomical structure of the human heart. The left ventricular outflow (LVO) is an anatomically complex region of the left ventricle (LV), which forms the ascending aorta (AA), supports the aortic valve leaflets, and houses important components of the conduction system. The left ventricle (LV) is located between the anterior leaflet of the mitral valve (MV) and the smooth side of the left ventricle of the muscular portion of the ventricular septum and the membranous interventricular septum (IVS). The anterior leaflet (posterior wall) of the mitral valve (MV) separates from the IVS by the left ventricular outflow (LVO). The region of fibrous continuity is called the mitral-aortic fibrous junction (MAIF).

[0026] The membranous septum (membranous IVS) is a small, compact fibrous portion of the IVS located between the non-coronary cusp and the right coronary cusp of the aortic valve and the septal cusp of the tricuspid valve (TV). Therefore, at its upper end, it is continuous with the right wall of the aortic root. The membranous septum is partially atrioventricular, anatomically separating a small portion of the right atrium closest to the septal cusp of the tricuspid valve (TV) from the left ventricle (LV). The atrioventricular (AV) bundle is located at the junction of the muscular portion of the ventricular septum and the membranous ventricular septum.

[0027] Figure 2 shows a schematic diagram of an embodiment of device 1 for segmenting membrane septa according to the present invention. Device 1 has a circuit 2, such as a processor or computer, configured to perform steps of method 10 for segmenting membrane septa, an embodiment of which is schematically shown in Figure 3.

[0028] The first step 11 of the method 10 is to segment the left ventricular outflow tract (LVOT) of the heart in a 3D image dataset, which may be obtained directly from an imaging system 3 (e.g., a CT imaging system, an X-ray system, etc.) or directly from an image storage or repository 4 (e.g., a hospital image archive or database). The 3D image dataset may be, for example, a cardiac CT angiography (CTA) image dataset obtained from a patient, which can be processed on the fly to provide important information to a physician, for example, in real time during surgery or to make a diagnosis.

[0029] The second step 12 of the method described above determines wall thickness information indicating the thickness of the septal wall at different locations in the segmented LVOT portions toward the right ventricle and right atrium. As described above, the wall thickness information is useful for identifying a membranous septum, as the septal thickness varies along the longitudinal extension of the septum.

[0030] The third step 13 maps the determined wall thickness information to the surface of the segmented LVOT, i.e., the surface of the segmented LVOT that leads to the right ventricle and right atrium, as used in step 12.

[0031] The fourth step 14 segments the membrane septum in the 3D image dataset based on the mapped wall thickness information. The membrane septum regions are then segmented by thresholding, independent component analysis, k-neighborhood (k-NN) clustering, or another method applied to the LVOT surface.

[0032] A fifth (optional) step 15 visualizes the wall thickness information and / or segmented results (segmented membrane septa) on the surface of the LVOT or on a properly flattened reformatted surface, which can be displayed, for example, on a screen 5. Furthermore, tools for the user to interactively modify the segmentation may be provided, for example, as a user interface.

[0033] A sixth (optional) step 16 allows the wall thickness information and / or segmentation results to be saved, for example, together with the 3D image dataset, or together with information indicating that it belongs to the 3D image dataset. For example, it can be saved as an annotation to the 3D image dataset.

[0034] In one embodiment, the device is implemented together with a dedicated unit or means, for example, an input unit for acquiring a 3D image dataset, a processing unit for performing the steps of method 10, and an output unit for outputting the results.

[0035] The input unit may be directly coupled to or connected to the imaging system, or it may acquire (i.e., retrieve or receive) these signals from storage, buffers, networks, or buses. Therefore, the input unit may be a (wired or wireless) communication interface or data interface, such as a Bluetooth® interface, Wi-Fi interface, LAN interface, HDMI interface, direct cable connection, or any other suitable interface that enables the transfer of signals to a device.

[0036] The processing unit may be configured to perform the steps of method 10, or it may be any kind of means configured to process a 3D image dataset. The processing unit can be implemented in software and / or hardware, for example, as a programmed processor or computer on a user device, or as an application.

[0037] The output unit is generally any interface that provides the determined result, for example, by transmitting the result to another device or providing the result for retrieval by another device (e.g., a computer, tablet, hospital network, etc.). Therefore, the output unit is generally any (wired or wireless) communication or data interface.

[0038] A more detailed embodiment of the method 20 according to the present invention, which uses model-based segmentation (MBS), is described below, as shown in the flowchart in Figure 4.

[0039] As the first step 21 of the segmentation method, segmentation of the heart and aortic valve is performed (generally including step 11 above, which segments the LVOT) using a model-based method, for example, as described in Waechter I. et al., “Patient Specific Models for Planning and Guidance of Minimally Invasive Aortic Valve Implantation” in “Medical Image Computing and Computer-Assisted Intervention - MICCAI 2010”, Lecture Notes in Computer Science 2010, vol 6361, 526 - 533. This method uses model-based segmentation to fit the image to a cardiac model that includes a detailed model of the aortic valve, as shown in Figure 5A (obtained from this disclosure). In addition, information is encoded in the model that enables, for example, the creation of a plane perpendicular to the outflow tract or the determination of the aortic valve annular surface, as shown in Figure 5B (also obtained from this disclosure). The information encoded in the model is supplemented with information about regions where membranous septa can be found (for example, by using information from publications or by using regions where membranous septa have been found in previous cases, expanded by a safety margin).

[0040] In the second step 22 (which is substantially identical to step 13), wall thickness information is mapped onto the outflow channel where a membrane septum is expected to be present. This can be done by the following substeps.

[0041] In the first substep 221, an axis approximating the centerline of the outflow tract is determined. This can be done by determining the aortic valve annular surface 30 (see Figure 6, which shows a cross-sectional view passing through the AA and LVOT) and defining an axis 31 perpendicular to this aortic valve annular surface 30 in the middle of the region where this aortic valve annular surface 30 intersects the outflow tract (LVOT). Alternatively, the axis can be defined by fitting a line passing through the centroids of two or more surface rings of the ascending aorta, in particular as defined in a adapted geometric mode. Figure 6A shows a cross-sectional view of the valve annular surface 30 shown in Figure 6B. Figure 6B shows a cross-sectional view perpendicular to the aortic valve annular surface 30 and passing through another plane 32 (the long axis plane for the aortic valve, as opposed to the long axis plane for the left ventricle) passing through axis 31. Line 33 indicates the wall of AA (in Figures 6A and 6B), and LV and LVOT (in Figure 6B).

[0042] In the second substep 222, using axis 31, a set of high-density rays (one ray located in another plane 32 in Figure 6A) is defined that are perpendicular to axis 31 and pass through the outflow tract wall 33 in a region where a membrane septum can be found according to the encoded label in the heart model. For each of these rays, a quantitative characteristic (alternative measure) of the thickness of the septum wall is calculated and mapped to the mesh surface.

[0043] In the third substep 223, an example of such quantity is, for example, the average gray value (or pixel value or image value) of a ray segment of a given length, starting from the surface of the LVOT. This average gray value decreases as there is more non-enhanced tissue between the angiogenic ventricles. Alternatively, machine learning can be used to train a classifier that classifies the shape as belonging to or not belonging to the septum, or to train a regression network that estimates the wall thickness.

[0044] In the third step 23, the membranous septum region is segmented, which can be done by thresholding, independent component analysis, k-NN clustering, or another method that uses a quantity characteristic of the wall thickness as a feature. Visualization of the segmentation results and interactive modification (e.g., by editing contours on a curved surface after appropriately transforming the segmentation results) can be done using standard methods, such as an electronic pen or brush, to draw or modify 2D contours on a 2D region.

[0045] Figures 7 to 11 show the method and its results. Figure 7 shows a cross-sectional view similar to the one shown in Figure 6B. Figure 7 shows the search space in the z direction (indicated by arrow 40) along the axis 31 from the aortic valve annular surface 30 to the left ventricular (LV) surface 34. Arrow 41 indicates the approximate location of the membranous septum. Figure 8 shows a cross-sectional view similar to the one shown in Figure 6, but with a different plane 35 rotated around the axis compared to the plane 32 shown in Figure 6A, cross-sectional view 8B is shown.

[0046] Figure 9 shows a cross-sectional view similar to the one shown in Figure 6A to illustrate how the rays (shown as ray 36 in Figure 9A and ray 37 in Figure 9B) rotate around axis 31. This is the search space in the angular direction (indicated by arrow 42).

[0047] Figure 10 shows the results of an MBS-based implementation that gives a polyhedron / map covered with translucent RV and RA as a texture on the mesh surface of the model. Figure 11 shows the polyhedron / map using the occupied regions of RV and RA.

[0048] Figure 12 illustrates another exemplary embodiment of the present invention. As shown in Figure 12A, rays are sampled from the centroid of a mesh triangle of a mesh modeling portion of the anatomical structure of the heart (including the right ventricle) in a direction perpendicular to the axis 31 of the LVOT. These rays may have a length in the range of 5 to 10 mm, for example, 6 or 10 mm, where a predetermined value is used, and this value may be set to a value that occupies a thin portion of the septum thickness. The starting point of the rays is generally on the LVOT / aortic plane. The average density (HU) of the rays is mapped to the rendered gray value color (image value or pixel value) of the triangular mesh. Thus, for example, light gray can indicate a low-density portion where the septum is thin, and dark gray can indicate a low-density portion where the septum is thick.

[0049] Figure 12B shows a flattened (i.e., curve-reformatted) two-dimensional polyhedron, parameterized by cylindrical mapping of the LVOT and aortic mesh surface using the height and angle of the LVOT axis. Sampling rays are densely sampled from the flattened two-dimensional polyhedron rather than originating from a few triangular centroids.

[0050] Figure 12C shows a diagram in which the right ventricle is obscured for improved visual inspection so as not to obstruct the septum. The densely sampled two-dimensional polyhedron shown in Figure 12B is used for texture mapping and folded around the triangular mesh representations of the LVOT and aortic surface to produce a finer, and therefore clearer, representation.

[0051] Instead of using model-based segmentation in step 21, a neural network can also be used to segment the outflow tract as a voxel mask. In substep 221, the axis used to project the ray can then be determined by calculating a distance map of the interior of the outflow tract and maximizing the straight-line distance within the outflow tract to the surface of this outflow tract. A rough region in which the membrane septum can be found is defined by segmenting the right ventricle and right atrium, finding the point closest to the LVOT, identifying the ray passing through this point, defining the angular range for rotating this ray around the axis, and the translational range for moving the ray above or below the axis. Identification of the membrane septum by projecting a ray and calculating a quantity indicating the wall thickness can be done using the MBS in substeps 222 and 223 as described in the above embodiment.

[0052] The construction of 2D polyhedra from triangular meshes and LVOT axes is performed as follows: For each cylindrical coordinate (z, Ф) on an LVOT axis, a ray orthogonal to the axis is projected from this axis, allowing identification of the nearest intersecting triangle vertex (e.g., performed as an efficient KD (K-dimensional) tree search). Then, the 3D coordinates of the surface points of each polyhedron are determined, for example, as the k-nearest neighbors average of the mesh vertices. Thus, all landmarks that can help define membranous septa can be correlated in all three representations (image volume, surface mesh, and reformatted polyhedron) for both manual annotation and automated algorithmic purposes.

[0053] In summary, the present invention enables the efficient and reliable segmentation of membrane septa in an automated manner. The disclosed method can be used in conjunction with other types of imaging, such as MR images or spectral CT images. Spectral CT images are used to better resolve the material properties of the underlying tissue.

[0054] The segmentation results can be used to visualize the membranous septum and extract measurements such as the distance between the aortic valve surface and the lower and / or upper edges of the septum in a medical diagnostic viewing system, for example, in a transcatheter aortic valve replacement (TAVR) application.

[0055] Although the present invention has been illustrated and described in detail in the drawings and the above description, such illustrations and descriptions should be considered illustrative or typical and not limiting. The present invention is not limited to the disclosed embodiments. Other variations of the disclosed embodiments can be understood and practiced by those skilled in the art in carrying out the claimed invention by considering the drawings, this disclosure and the appended claims.

[0056] In a claim, the term “having” does not preclude other elements or steps, nor does it preclude multiple elements or steps if it is not stated that there are multiple. A single element or other unit may fulfill the functions of several items enumerated in a claim. The mere fact that certain means are described in different dependent claims does not imply that combinations of these means cannot be used advantageously.

[0057] Computer programs may be stored / distributed on suitable non-transient media, such as optical storage media or solid-state media, supplied together with or as part of other hardware, but they may also be distributed in other forms, such as via the Internet or other wired or wireless communication systems.

[0058] No reference numeral in a claim should be construed as limiting its scope.

Claims

1. A device for segmenting membrane septa in a 3D image dataset, In the aforementioned 3D image dataset, the left ventricular outflow tract of the heart is segmented. The wall thickness information indicating the thickness of the membranous septum wall at different locations in the portion of the segmented left ventricular outflow tract toward the right ventricle and right atrium is determined. The determined wall thickness information is mapped onto the surface of the portion of the segmented left ventricular outflow tract, and Based on the mapped wall thickness information, the membrane septum is segmented in the 3D image dataset. A device having a circuit configured to meet certain requirements.

2. The device according to claim 1, wherein the circuit is further configured to segment the membrane septum by applying one of the following to the surface of the segmented portion of the left ventricular outflow tract: thresholding, independent component analysis, and k-neighboring clustering.

3. The circuit is further configured to segment the left ventricular outflow tract using model-based segmentation in order to fit the cardiac model to the 3D image dataset, in particular, Having an index of the region in which the membranous septum should be found, and / or The information includes the generation of two or more planes perpendicular to the left ventricular outflow tract and / or the determination of the aortic valve annular surface. The device according to claim 1 or 2, configured to use a cardiac model.

4. The aforementioned circuit is Determining an axis that approximates the centerline of the left ventricular outflow tract, Defining a series of rays perpendicular to the axis and / or the surface, passing through the wall of the left ventricular outflow tract in the portion of the segmented left ventricular outflow tract, To determine the corresponding wall thickness information for the aforementioned series of light rays, The wall thickness information is mapped to the position of the surface on which each of the aforementioned light rays passes through the wall of the left ventricular outflow tract. The device according to any one of claims 1 to 3, further configured to determine the wall thickness information by...

5. The aforementioned circuit is Before determining the wall thickness information, the surface of the aortic valve annularity is determined, Defining an axis perpendicular to the aortic valve annular surface and The device according to claim 4, further configured to determine the approximating axis by means of the device.

6. The device according to claim 4, wherein the circuit is further configured to determine the axis by fitting a line passing through the centroids of two or more surface rings of the ascending aorta, in particular, as defined by a fitted geometric model.

7. The device according to claim 4, wherein the circuit is further configured to determine the wall thickness value of the septum along each of the light rays, or an alternative value of the wall thickness value, as wall thickness information.

8. The aforementioned circuit is As a substitute value for the wall thickness, determine the average image value of each ray segment of the light beam, starting from the surface of the left ventricular outflow tract and having a predetermined or adaptable length, and / or The wall thickness value is determined based on the length of each ray segment of the light beam, starting from the surface of the left ventricular outflow tract and ending at a change in the image value indicating a change in tissue. The device according to claim 7, further configured as follows.

9. The device according to claim 7, wherein the circuit is further configured to determine the wall thickness information by using a trained algorithm or computer system, in particular trained machine learning, a trained classifier, or a trained neural network, which has been trained to estimate the thickness of the walls of organic structures in a 3D image dataset of the heart.

10. The device according to claim 4, wherein the circuit is further configured to calculate a distance map of the interior of the left ventricular outflow tract and to determine the axis by maximizing the straight-line distance within the left ventricular outflow tract to the surface of the left ventricular outflow tract.

11. The aforementioned circuit is Segmenting the right ventricle and right atrium, Finding the point closest to the left ventricular outflow tract, Identifying the aforementioned ray passing through this point, To define an angular range for rotating the light ray around the axis, and a translational range for moving the light ray above or below the axis. The device according to claim 4, further configured to determine the region in which the membranous septum should be found.

12. The aforementioned circuit is Using a geometric model represented by a triangular mesh, for each outward-facing ray from the surface of the polyhedron of the left ventricular outflow tract, the 3D coordinates at which the ray intersects with one of the mesh triangles intersecting the vertices of the triangles are identified, and The 3D coordinates are used as the starting point of the ray for estimating the local thickness of the septum. The device according to claim 4, further configured as follows.

13. The device according to any one of claims 1 to 12, wherein the circuit is further configured to visualize the determined wall thickness information and / or the segmented membrane septum on the surface of the portion of the segmented left ventricular outflow tract, or on the flattened reformatting of the portion of the segmented left ventricular outflow tract.

14. A computer implementation method for segmenting membranous septa in a 3D image dataset, Segmenting the left ventricular outflow tract of the heart in a 3D image dataset, Determining wall thickness information indicating the wall thickness of the membranous septum at different locations in the segmented left ventricular outflow tract, specifically in the portion leading to the right ventricle and right atrium, The determined wall thickness information is mapped onto the surface of the portion of the segmented left ventricular outflow tract. Based on the mapped wall thickness information, the membrane septum is segmented in the 3D image dataset. A computer implementation method having

15. A computer program having program code means, wherein the program code means causes the computer to perform the steps of the method according to claim 14 when the computer program is executed on the computer.

Images