Tumor ablation planning method for optimizing tumor point clouds, tumor ablation planning method, tumor point cloud optimization system, and tumor ablation planning system

Abstract

To optimize the number of tumor point clusters used in clustering, thereby significantly improving clustering efficiency while ensuring coverage. [Solution] The tumor point cloud optimization method according to the embodiment is a tumor point cloud optimization method in tumor ablation planning that obtains an optimized tumor point cloud for use in tumor ablation planning by optimizing a tumor point cloud representing a collection of three-dimensional voxels of a tumor, and includes an acquisition step of acquiring the tumor point cloud and dividing the tumor point cloud into internal points representing a collection of three-dimensional voxels inside the tumor and surface points representing a collection of three-dimensional voxels on the tumor surface; an internal sampling step of downsampling the internal points to obtain internal sampling points; a surface sampling step of performing sampling on the surface points different from that in the internal sampling step to obtain surface sampling points; and an output step of outputting the internal sampling points and the surface sampling points as the optimized tumor point cloud.

Description

[Technical Field]

[0001] Embodiments disclosed herein and in the drawings relate to a tumor point cloud optimization method, a tumor ablation planning method, a tumor point cloud optimization system, and a tumor ablation planning system in tumor ablation planning. [Background technology]

[0002] Ablation is a treatment method that destroys abnormal tissue or cells by physical or chemical means to treat disease. Among these, RFA (Radiofrequency Ablation), MWA (Microwave Ablation), and Cryoablation are currently effective methods used for minimally invasive treatment of liver tumors, kidney tumors, and other conditions. To ensure the safety and effectiveness of ablation treatment, it is necessary to accurately plan the ablation, determining the size, shape, and location of the ablation area. Ablation planning is a process that automatically plans ablation treatment using computational technology and medical imaging data. Performing automated ablation planning effectively and efficiently based on the shape of the tumor is extremely important for ablation treatment, and how to support physicians in automatically planning ablation is of great significance in ablation treatment.

[0003] Ablation planning typically relies on highly accurate patient anatomical data. Point clouds are a common representation of medical anatomical data applied to tissues such as tumors. Point clouds are datasets consisting of numerous discrete points that accurately represent the surface and volume of an object in three-dimensional space, and are usually derived from medical imaging data such as computed tomography (CT) or magnetic resonance imaging (MRI) scans. The efficiency of an ablation plan is clearly negatively correlated with the number of tumor voxels and the point cloud distribution of the tumor. That is, the more tumor voxels there are and the denser the point cloud distribution of the tumor, the less efficient the ablation plan becomes.

[0004] However, conventional algorithms for automated tumor ablation planning based on point clouds either ignore the shape distribution of the tumor or ignore the impact of the number of tumor point clouds on efficiency.

[0005] For example, Patent Document 1 discloses a method for accurately simulating radio frequency ablation technology by covering a tumor with an ellipsoid. In this method, the covariance matrix of the tumor point cloud is used to establish a relationship between the ablation size and the clustering distance metric. Specifically, first, the covariance matrix is ​​modified based on the length of the major and minor semi-axis of the ellipsoid defined by the user. Next, the tumor point cloud data is clustered using the Mahalanobis distance defined by the covariance matrix. Then, each cluster is covered with the size of the ablation ellipsoid set by the user. Here, if all clusters are covered by ablation, the number of ablations is output as the minimum value; otherwise, 1 is added to the clustering number and the above procedure is repeated.

[0006] The method described in Patent Document 1 involves many iterative processes for clustering tumor point clusters. As shown in Table 1, the inventors' test data revealed that the method in Patent Document 1 is efficient for planning tumors with small volumes (e.g., 55 mL or less), but for tumors with large volumes (e.g., exceeding 67.3 mL), the number of iterations increases significantly, and good efficiency cannot be obtained. Table 1 shows a comparison of the number of iterations for the method in Patent Document 1.

[0007] [Table 1]

[0008] The main reason for the above problem is that larger tumors contain more voxels, requiring more time and iterations for clustering. To solve this problem, downsampling can be used to improve the efficiency of clustering.

[0009] However, downsampling also introduces new problems. For example, using uniform downsampling can lead to the loss of important information in the edge point cloud of the tumor surface, potentially preventing ablation from fully covering these points. Geometric sampling places more emphasis on geometric abrupt changes on curved surfaces and avoids the information loss associated with uniform downsampling, but at the cost of non-uniform sampling results, which can lead to an increase in the number of ablations. [Prior art documents] [Patent Documents]

[0010] [Patent Document 1] Chinese Patent Application Publication No. 105997245 Specification [Patent Document 2] Special Publication No. 2010-506621 [Patent Document 3] Special Publication No. 2016-135177 [Non-patent literature]

[0011] [Non-Patent Document 1] "Point cloud simplification with preserved edge based on normal vector" Huiyan Han, Xie Han, Fusheng Sun, Chunyan Huang(https: / / www.sciencedirect.com / science / article / abs / pii / S0030402615004052) [Overview of the Initiative] [Problems that the invention aims to solve]

[0012] One of the problems that the embodiments disclosed herein and in the drawings aim to solve is to optimize the number of tumor point clusters used for clustering and to significantly improve the efficiency of clustering while ensuring coverage. However, the problems that the embodiments disclosed herein and in the drawings aim to solve are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0013] The tumor point cloud optimization method according to the embodiment is a tumor point cloud optimization method in tumor ablation planning that obtains an optimized tumor point cloud for use in tumor ablation planning by optimizing a tumor point cloud representing a collection of three-dimensional voxels of a tumor, and includes an acquisition step of acquiring the tumor point cloud and dividing the tumor point cloud into internal points representing a collection of three-dimensional voxels inside the tumor and surface points representing a collection of three-dimensional voxels on the tumor surface; an internal sampling step of downsampling the internal points to obtain internal sampling points; a surface sampling step of performing sampling on the surface points different from that performed in the internal sampling step to obtain surface sampling points; and an output step of outputting the internal sampling points and the surface sampling points as the optimized tumor point cloud. [Brief explanation of the drawing]

[0014] [Figure 1] FIG. 1 is a schematic diagram for explaining the principle of the tumor point cloud optimization method according to the embodiment. [Figure 2] FIG. 2 is a flowchart showing an example of the tumor point cloud optimization method according to the embodiment. [Figure 3] FIG. 3 is a schematic diagram for explaining a specific aspect of dividing the tumor point cloud into surface points and internal points according to the embodiment. [Figure 4] FIG. 4 is a schematic diagram for explaining rasterization processing performed on surface points according to the embodiment. [Figure 5] FIG. 5 is a schematic diagram showing test results of testing an ablation plan for a liver tumor according to the first embodiment. [Figure 6] FIG. 6 is a schematic diagram showing an example of a user interface screen of an ablation plan according to the embodiment. [Figure 7] FIG. 7 is a flowchart showing a tumor point cloud optimization method according to a second embodiment. [Figure 8] FIG. 8 is a flowchart showing a tumor point cloud optimization method according to a third embodiment. [Figure 9] FIG. 9 is a block diagram showing an example of a tumor point cloud optimization system according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0015] Hereinafter, with reference to the accompanying drawings, embodiments of a tumor point cloud optimization method, a tumor ablation planning method, a tumor point cloud optimization system, and a tumor ablation planning system in tumor ablation planning according to embodiments will be described in detail. Note that the tumor point cloud optimization method, tumor ablation planning method, tumor point cloud optimization system, and tumor ablation planning system in tumor ablation planning according to the present invention are not limited to the embodiments shown below. In the following description, the same components are denoted by the same reference numerals, and overlapping descriptions are omitted.

[0016] (First Embodiment) First, with reference to Figure 1, the principle of the tumor point cloud optimization method according to the first embodiment will be explained.

[0017] Figure 1 is a schematic diagram illustrating the principle of the tumor point cloud optimization method according to this embodiment.

[0018] As shown in Figure 1, the key to the tumor point cloud optimization method according to this embodiment is to divide the tumor point cloud into internal points, which represent a collection of three-dimensional voxels inside the tumor, and surface points, which represent a collection of three-dimensional voxels on the tumor surface, and to perform different downsampling on the internal points and surface points.

[0019] In the clustering process, internal points overlap significantly; therefore, in this embodiment, internal points are sampled at a lower sampling density than surface points in order to obtain internal sampling points. In this embodiment, uniform downsampling is employed for internal points. On the other hand, since surface points mainly reflect tumor shape and distribution, in this embodiment, downsampling is performed on surface points using a weighted grid voxel method to obtain surface sampling points. This allows for obtaining more keypoints to be used in the clustering process and ensures coverage. In this embodiment, the combination of internal and surface sampling points is used as an optimized tumor point set for planning tumor ablation.

[0020] Therefore, the tumor point cloud optimization method according to this embodiment obtains an optimized tumor point cloud for use in tumor ablation planning by optimizing the tumor point cloud, which represents a collection of three-dimensional voxels of the tumor. The tumor point cloud optimization method includes an acquisition step of acquiring the tumor point cloud and dividing the tumor point cloud into internal points, which represent a collection of three-dimensional voxels inside the tumor, and surface points, which represent a collection of three-dimensional voxels on the tumor surface; an internal sampling step of downsampling the internal points to obtain internal sampling points; a surface sampling step of performing sampling on the surface points different from that performed in the internal sampling step to obtain surface sampling points; and an output step of outputting the internal sampling points and surface sampling points as an optimized tumor point cloud.

[0021] The following describes specific embodiments of the tumor point cloud optimization method. Figure 2 is a flowchart showing an example of the tumor point cloud optimization method according to this embodiment.

[0022] As shown in Figure 2, the tumor point cloud optimization method according to this embodiment first involves inputting the acquired tumor point cloud.

[0023] A tumor point cloud consists of numerous discrete points, each representing a coordinate on or within the tumor surface. Acquisition of a tumor point cloud can be achieved by techniques well known in the art. For example, first, three-dimensional tumor data of a patient can be acquired using medical imaging techniques (e.g., CT, MRI). Next, preprocessing, including denoising, contrast enhancement, and segmentation, can be performed on the acquired image data to more accurately identify the tumor region. Finally, point cloud data of the tumor region can be extracted from the preprocessed medical images.

[0024] In step 100, the tumor point group is divided into internal points and surface points. The specific method for dividing the tumor point group into internal and surface points will be described in detail later.

[0025] Next, in step 200, an internal sampling step is performed to obtain internal sampling points by uniformly downsampling the internal points.

[0026] Next, in steps 300 to 500, a surface sampling step is performed, and surface sampling points are obtained by performing sampling on surface points differently from that in the internal sampling step.

[0027] As described above, in this embodiment, surface points are sampled using a so-called weighted grid voxel method, taking advantage of the fact that surface points mainly reflect the shape and distribution of the tumor point group. Specifically, in step 300, the surface points are first rasterized, then in step 400, the centroid of each surface point in each grid is calculated, and then in step 500, the sampling result of the surface points is obtained based on the distance from the centroid of each surface point in the grid and the weight of the number of surface points in the grid.

[0028] Finally, in step 600, the sampling results of internal points and surface points are integrated and output as the sampling result of the tumor point cluster.

[0029] The specific details of each step in the flowchart shown in Figure 2 will be explained below with reference to the diagram.

[0030] Figure 3 is a schematic diagram illustrating a specific method for dividing the tumor point group in step 100 into surface points and internal points.

[0031] As shown in Figure 3(a), first, a 3D bounding box that can include the entire tumor is calculated from the tumor point cloud data.

[0032] Next, the point cloud data within the bounding box is divided into tumor and non-tumor regions. This division process can be achieved using known image processing techniques such as threshold segmentation, region expansion, and machine learning classification, and since these techniques are well known to those skilled in the art, their explanation is omitted here.

[0033] Next, the bounding box interior region is examined along the X, Y, and Z axes, respectively, and edge jump points between the obtained tumor region and non-tumor region are detected. The set of these edge jump points is defined as the surface points of the tumor point group, and the points other than the surface points are defined as the interior points of the tumor point group. An edge jump point refers to a discontinuous or abruptly changing point in edge detection. Edge jump point detection can be achieved by known general edge detection algorithms. For example, edge jump point detection can be achieved by the Sobel operator, the Canny edge detection algorithm, the zero-crossing algorithm, or other machine learning-based edge detection methods. These techniques are well known to those skilled in the art, so their explanation is omitted here.

[0034] In Figure 3(b), the colored areas (areas indicated by the diagonal pattern) are examples of tumor regions in one slice along the Z axis, and the black dots are examples of obtained edge jump points, i.e., surface points.

[0035] The method for separating surface points from interior points is not limited to this, and as is known to those skilled in the art, it can also be achieved by other image processing or machine learning algorithms. For example, it can be achieved by surface reconstruction algorithms, distance-based methods, geometric feature-based methods, point cloud density analysis, normal vector analysis, and region expansion algorithms.

[0036] Furthermore, to obtain a more accurate surface model, the extracted surface points may be further smoothed and optimized to reduce the effects of noise, mis-segmentation, and misclassification.

[0037] The following describes a specific manner in which the internal sampling step in step 200 is performed.

[0038] According to this embodiment, internal points are sampled by uniform downsampling. Uniform downsampling is a method of reducing the amount of point cloud data and is achieved by uniformly selecting some points from the original point cloud.

[0039] Specifically, in this embodiment, internal sampling points are obtained by uniformly downsampling the internal points obtained in step 100 using the algorithm shown in equation (1).

[0040]

number

[0041] In equation (1), Q is the sampling set of points inside the tumor, p n is each internal point within the tumor point group. N is the number of points within the internal point group (n ≤ N), a is the index value of the point in the point group, and step is the sampling step representing the interval between sampling points.

[0042] Equation (1) represents only one aspect of uniform downsampling, and the embodiments are not limited thereto. Uniform downsampling of internal points may be performed using other algorithms well known to those skilled in the art, such as uniform grid downsampling, spatial pyramid downsampling, minimum distance downsampling, uniform radius downsampling, or octree downsampling. An appropriate downsampling method can be selected depending on the specific application and requirements for the point cloud data, but such a description is omitted here.

[0043] This embodiment reduces the complexity of data processing with a relatively simple algorithm while maintaining the overall structure and shape characteristics of the point cloud by uniformly downsampling internal points.

[0044] The following describes a specific method for performing weighted grid voxel downsampling on surface points in steps 300 to 500.

[0045] The weighted grid voxel method according to this embodiment is a grid-based weighted downsampling method, which includes a rasterization step 300 in which surface points are rasterized using a cubic grid composed of multiple blocks to obtain surface blocks that are blocks corresponding to each point among the surface points, wherein the size of each block corresponds to a point in the tumor point group; a centroid calculation step 400 in which the centroid position of each surface block in the cubic grid is calculated for each cubic grid; and a sampling block selection step 500 in which the distance of each surface block from the calculated centroid position is calculated for each cubic grid, and a predetermined number or less of surface blocks are weighted and selected as sampling blocks from among the surface blocks based on the calculated distance and the weight related to that distance. The surface points corresponding to the sampling blocks are used as the sampling points of the surface points.

[0046] Figure 4 is a schematic diagram illustrating the rasterization process performed on surface points in step 300.

[0047] As shown in Figure 4, first, the minimum 3D bounding box of the tumor surface points is calculated, and then the surface of the 3D bounding box is rasterized with a cubic grid of side length S to obtain a set of cubic grids G and the number of cubic grids I.

[0048] Next, the center G of each cube grid i k neighborhoods (k=S 3 -1) is examined, and the number L of surface blocks, which are blocks containing tumor surface points T in each cubic grid, is determined. The right side of Figure 4 shows a schematic diagram of surface blocks in a cubic grid that has been rasterized on the surface of a 3D bounding box. The colored blocks (gray areas) in this cubic grid represent the presence of surface points among the tumor voxel points in that block, and these colored blocks are referred to as surface blocks.

[0049] Next, the centroid M of the surface block in each cubic grid. iis calculated by the following formula (2). In the case of "L=0", that is, when this cubic grid does not contain any surface blocks, the calculation of the centroid of this cubic grid is omitted.

[0050]

Mathematical Expression

[0051] Next, for a cubic grid whose centroid has been calculated, the distance from the obtained centroid to each surface block in each cubic grid is calculated. In the present embodiment, the distance is calculated by the following formula (3) using Manhattan distance. Here, x l , y l , z l are the 3D coordinates of each surface block in the cubic grid, x m , y m , z m are the 3D coordinates of the obtained centroid.

[0052]

Mathematical Expression

[0053] Next, in each cubic grid, downsampling is performed by sorting each surface block according to the distance from the centroid, and selecting surface blocks according to the sorting result.

[0054] Specifically, the surface blocks in each cubic grid are downsampled by the following formula (4). Here, P i is the result obtained after downsampling the surface blocks. D j is a set of surface blocks having the same distance from the centroid, J is the number of sets of surface blocks having different distances from the centroid, ω j is the selection weight for a surface block, and when the surface block is sampled, ω j is "1", and otherwise it is "0".

[0055]

Mathematical Expression

[0056] D j The number of surface blocks in this embodiment is not necessarily limited to one. According to this embodiment, if there are multiple surface blocks with the same distance from the center of gravity, the more surface blocks there are, the greater the selection weight given to those surface blocks at the same distance (hereinafter referred to as "candidate surface blocks").

[0057] In this embodiment, the selection weights for candidate surface blocks that are the same distance from the centroid are determined by the following algorithm.

[0058] First, starting from the center of gravity, D j From the relationships between the vectors drawn on each candidate surface block inside (hereinafter referred to as "candidate vectors"), D j D' j This is simplified to the following: Specifically, for example, an arbitrary candidate vector is used as the reference candidate vector, and the relationship between the angles between other candidate vectors and the reference candidate vector is used to determine D j This can be simplified. For example, a candidate vector angle threshold can be set for the angle between candidate vectors. If the angle between two candidate vectors is relatively large and greater than the candidate vector angle threshold, both surface blocks corresponding to these two candidate vectors are sampled; otherwise, one of them is discarded. The above procedure is repeated until all candidate vectors have been compared and selected.

[0059] In this embodiment, the candidate vector angle threshold can be set according to the step used for downsampling internal points. For example, the candidate vector angle threshold can be set to be proportional to the step. For instance, if the step is 3, the angle threshold can be set to 30 degrees, and if the sampling step is 6, the angle threshold can be adjusted to 60 degrees. By setting the candidate vector angle threshold to be proportional to the number of steps in the downsampling of internal points, a better balance can be achieved between the sampling ratio of internal and external points.

[0060] As a result, the simplified equation (4) above becomes equation (5) below.

[0061]

number

[0062] The downsampling described above is performed on each surface block within one cubic grid. The downsampling results for the one cubic grid are then cumulatively added to obtain the surface point sampling results shown in equation (6) below.

[0063]

number

[0064] Next, in step 600, the combination of internal sampling points obtained in step 200 and surface sampling points obtained in step 500 is output as an optimized tumor point cloud.

[0065] According to the first embodiment, the tumor point cloud is divided into internal points and surface points, and different sampling methods are used for the internal and surface points to obtain internal and surface sampled points, and the combination of internal and surface sampled points is output as an optimized tumor point cloud. In particular, this embodiment samples surface points using grid-weighted downsampling. By using the surface sampled points obtained in this way as an optimized tumor point cloud for tumor ablation planning, this embodiment can optimize the number of tumor point clouds used for clustering, and can significantly improve the efficiency of clustering while guaranteeing ablation coverage.

[0066] Based on the technical proposal of the first embodiment described above, the inventor conducted tests on ablation plans for liver tumors and evaluated the test results. The evaluation results are shown in Figure 5.

[0067] Figure 5 is a schematic diagram showing the test results of testing a liver tumor ablation plan according to the first embodiment.

[0068] The left side of Figure 5 shows the results of ablation planning based on all tumor point clusters, and the right side of Figure 5 shows the results of ablation planning based on the optimized tumor point cluster of the first embodiment. As is clear from Figure 5, the size and distribution of the ablation ellipsoids obtained by the first embodiment based on the optimized tumor point cluster are very similar to those obtained by the tumor ablation planning based on all tumor point clusters, and the coverage and clustering efficiency of this embodiment are clearly superior to the prior art.

[0069] In the calculation of the first embodiment, an example of a cubic grid with side length "S=3", i.e., "3×3", was shown, but the side length S in this embodiment is not limited to this. The size of the cubic grid can be appropriately adjusted according to the size of the tumor and the size of the minimum 3D bounding box of the tumor surface points calculated in step 300.

[0070] The sampling step step in step 200, which uniformly downsamples points within the tumor, may be set appropriately depending on the size of the tumor.

[0071] These setting functions can be incorporated into the user interface screen, for example, as shown in Figure 6. Figure 6 is a schematic diagram showing an example of the user interface screen for ablation planning according to this embodiment. This embodiment can also provide the user with a function to display comparison results when different setting values ​​are entered via the user interface screen.

[0072] The tumor point cloud optimization method according to this embodiment may be performed in a medical imaging diagnostic device such as an ultrasound diagnostic device, a CT imaging device, or an MRI imaging device, or it may be performed in an independent device such as a workstation.

[0073] (Second embodiment) Although a first embodiment of the tumor point cloud optimization method has been described above, the embodiment is not limited thereto. For example, the tumor point cloud optimization method according to the first embodiment may be combined with a conventional manual ablation planning method.

[0074] The second embodiment is a tumor point cloud optimization method that combines a tumor point cloud optimization method with a conventional manual ablation planning method, and the second embodiment will be described in detail below with reference to Figure 7.

[0075] The description of the second embodiment will mainly explain the differences from the first embodiment described above. In the description of the second embodiment, components identical to those in the first embodiment described above will be denoted by the same reference numerals and their descriptions will be omitted.

[0076] Figure 7 is a flowchart illustrating the tumor point cloud optimization method according to the second embodiment. Steps S10 and S20, enclosed in dashed boxes, are newly added steps. Steps 100' and 600' are different from those in the first embodiment.

[0077] In step S10, the tumor point cloud optimization method according to this embodiment first performs a user-manual ablation plan on the input tumor point cloud and retains the plan results.

[0078] Specifically, based on the location and orientation of the manual ablation plan, the ablation coverage can be obtained using the following equation (7).

[0079]

number

[0080] In the formula, C and C' are the foci of the ellipsoidal ablation, which are manually specified by the user. t is the direction of ablation, O is the center of ablation, and a is the semi-major axis.

[0081] Next, in step S20, the ablation plan obtained in step S10 is evaluated, and the uncovered regions in the plan results are obtained. The points included in the uncovered regions are designated as the point cloud required for automatic clustering in subsequent steps.

[0082] Specifically, in step S20, points in the tumor point cluster that are covered by the ablation range that the user manually planned in step S10 are excluded, and the remaining points are used as the point cluster for automated clustering.

[0083] Next, in step 100' of this embodiment, the point cloud obtained in step S20, which is necessary for automatic clustering, is divided into surface points and interior points, and the subsequent steps are continued.

[0084] In step S10, the user-manual ablation plan may generate one ablation range or multiple ablation ranges. In the latter case, in step S20, points covered by these multiple ablation ranges are excluded, and the remaining points are used as the point cloud required for automatic clustering.

[0085] In other words, the tumor point cloud optimization method according to this embodiment further includes a step in which the user pre-plans to set one or more specified ablation ranges in the tumor point cloud, and in the acquisition step, tumor point clouds outside the one or more specified ablation ranges are acquired and divided into internal points and surface points.

[0086] In step 600', the one or more points within the specified ablation range obtained by the user in step S10, the internal sampling points obtained in step 200, and the surface sampling points obtained in step 500 are output as an optimized tumor point cluster.

[0087] Other steps of the tumor point cloud optimization process according to the second embodiment are the same as those in the first embodiment, and therefore a detailed description is omitted here.

[0088] The semi-automatic mode according to the second embodiment offers improved practicality in clinical practice because it allows users to easily participate in the automated ablation plan. Manually set ablation not only covers areas set by the physician based on clinical experience, but more importantly, the pre-planned direction and position based on manual ablation is useful for direction planning in the automated ablation plan, for example, by avoiding important bones, blood vessels, and organs, which is extremely important in ablation planning. This mode can also be applied to situations where the user plans ablation for areas not covered by the automated ablation plan using a pre-made manual ablation plan, achieving the further effect of planning ablation for areas not covered by the automated ablation plan.

[0089] According to the second embodiment, by combining the tumor point cloud optimization method according to the embodiment with a conventional manual planning method, the efficiency of clustering can be significantly improved while ensuring ablation coverage, similar to the first embodiment, the number of tumor point clouds used for clustering can be further optimized, manual pre-planning can support directional planning in the automated ablation plan, and ablation plans can be realized for points that cannot be covered by the automated ablation plan.

[0090] (Third embodiment) The first and second embodiments of the tumor point cloud optimization method have been described above, but the method is not limited thereto. For example, in addition to the first and second embodiments, a mode for specifying the number of ablation procedures may be added.

[0091] The third embodiment adds a mode for specifying the number of ablation operations to the tumor point group optimization method, and the third embodiment will be described in detail below with reference to Figure 8.

[0092] The description of the third embodiment will mainly explain the differences from the first and second embodiments described above. In the description of the third embodiment, components identical to those in the first and second embodiments described above will be denoted by the same reference numerals and their descriptions will be omitted.

[0093] Figure 8 is a flowchart illustrating the tumor point cloud optimization method according to the third embodiment. The steps enclosed in the dashed box are the same as those in the first or second embodiment.

[0094] In step S1, the tumor point cloud optimization method according to this embodiment first involves the user specifying the number of ablation procedures.

[0095] In step 700, clustering is performed for the number of ablation cycles specified by the user.

[0096] According to this embodiment, if the clustering of tumor points is successful with the number of ablation procedures specified by the user, the ablation results are displayed; otherwise, relevant information is presented, and the user can modify the number of ablation procedures and repeat the clustering.

[0097] Other steps of the tumor point cloud optimization process according to the third embodiment are the same as those of the first or second embodiment, and therefore a detailed description is omitted here.

[0098] When executing an ablation plan, users have a need to consider and compare the results of clustering with different numbers of ablation cycles by specifying the number of ablation cycles. In this embodiment, to meet this need, a mode is provided that allows the user to specify the number of ablation cycles, thereby facilitating the user to quickly compare the different effects of different numbers of ablation cycles.

[0099] According to this embodiment, by providing a mode that allows the user to specify the number of ablation cycles, it is possible to significantly improve the efficiency of clustering while guaranteeing ablation coverage, similar to the first and second embodiments, and to optimize the number of tumor point groups used for clustering. Furthermore, by providing the user with a function to specify the number of clustering cycles, ablation coverage can be further improved by performing multiple clustering cycles.

[0100] (Other embodiments) In addition to the tumor point cloud optimization methods described in each of the embodiments above, it can also be implemented as a tumor point cloud optimization system.

[0101] The following describes an embodiment of the tumor point cloud optimization system. The tumor point cloud optimization system according to this embodiment optimizes the tumor point cloud, which represents a collection of three-dimensional voxels of a tumor, to obtain an optimized tumor point cloud for use in tumor ablation planning.

[0102] Figure 9 is a block diagram showing an example of a tumor point cloud optimization system according to this embodiment.

[0103] As shown in Figure 9, the tumor point cloud optimization system 10 according to this embodiment includes an acquisition device 1 that acquires a tumor point cloud and divides the tumor point cloud into internal points representing a collection of three-dimensional voxels inside the tumor and surface points representing a collection of three-dimensional voxels on the tumor surface; an internal sampling device 2 that downsamples the internal point cloud to obtain internal sampling points; a surface sampling device 3 that performs sampling on the surface points differently from that of the internal sampling device 2 to obtain surface sampling points; and an output device 4 that outputs the internal sampling points and surface sampling points as an optimized tumor point cloud.

[0104] In this embodiment, the acquisition device 1, internal sampling device 2, surface sampling device 3, and output device 4 can optimize the number of tumor point clusters used for clustering by performing each step described in the first to third embodiments above, thereby significantly improving the efficiency of clustering while ensuring ablation coverage. The acquisition device 1 is an example of an acquisition unit. The internal sampling device 2 is an example of an internal sampling unit. The surface sampling device 3 is an example of a surface sampling unit. The output device 4 is an example of an output unit.

[0105] Furthermore, this embodiment can also be implemented as a method and system for performing ablation planning using an optimized tumor point cloud obtained by a tumor point cloud optimization method or system.

[0106] In other words, the tumor ablation planning method according to this embodiment is a method for planning a tumor ablation based on a tumor point cloud representing a collection of three-dimensional voxels of a tumor, and is characterized by comprising the steps of: optimizing the tumor point cloud using the above-described tumor point cloud optimization method to obtain an optimized tumor point cloud; and planning the ablation of the tumor using the optimized tumor point cloud.

[0107] The tumor ablation planning system according to this embodiment is a system that performs tumor ablation planning based on a tumor point cloud representing a collection of three-dimensional voxels of a tumor, and is characterized by comprising a tumor point cloud optimization system that optimizes the tumor point cloud and obtains an optimized tumor point cloud, and a planning device that performs tumor ablation planning using the optimized tumor point cloud. Such a planning device is an example of a planning unit.

[0108] In the embodiments described above, uniform downsampling is employed for internal points, but the embodiments are not limited to these, and internal points may be sampled using other downsampling methods other than uniform downsampling.

[0109] Furthermore, in each of the embodiments described above, the sampling ratio for downsampling internal points can be set according to the actual situation in order to balance the distribution of tumor point groups. In this embodiment, the sampling rate for downsampling internal points is set based on the side length S of the cubic grid used for weighted downsampling of surface points. For example, the larger the side length S of the cubic grid, the smaller the sampling rate for internal points becomes; that is, the sampling rate for internal points may be inversely proportional to S. For example, when performing weighted downsampling using a cubic grid with "S=3" for surface points, the sampling rate for internal points is preferably "1 / (3*3*3)=1 / 27". That is, the sampling rate for internal points is preferably "P=1 / S3". From the test results, it was found that the effect of such a sampling rate for internal points is particularly good for tumors with a volume of "60 ml" or less.

[0110] The image processing, clustering, and other methods described in each of the embodiments above can all be implemented using various methods in the prior art, so a detailed explanation is omitted here.

[0111] The embodiments can be implemented not only as the tumor point cloud optimization method, system, and tumor ablation planning method, system described above, but also as a computer program product including a tumor point cloud optimization program and a tumor ablation planning program, or as a medium storing the tumor point cloud optimization program and the tumor ablation planning program.

[0112] The tumor point cloud optimization system and tumor ablation planning system according to the embodiment may be incorporated into a medical imaging diagnostic device, or the tumor point cloud optimization system and tumor ablation planning system may perform processing independently. In the latter case, the tumor point cloud optimization system and tumor ablation planning system have a processing circuit that performs processing similar to the steps described above, and a memory that stores programs corresponding to each function and various information. The processing circuit then acquires the tumor point cloud from a medical imaging diagnostic device such as an ultrasound diagnostic device or an image storage device via a network, and performs the above processing using the acquired tumor point cloud. Here, the processing circuit is a processor that realizes the function corresponding to each program by reading and executing a program from the memory.

[0113] In the above description, the term "processor" refers to circuits such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an Application Specific Integrated Circuit (ASIC), or a programmable logic device (e.g., a Simple Programmable Logic Device (SPLD), a Complex Programmable Logic Device (CPLD), and a Field Programmable Gate Array (FPGA)). A processor performs its functions by reading and executing a program stored in memory. Alternatively, instead of storing the program in memory, the processor may be configured to directly incorporate the program into its circuitry. In this case, the processor performs its functions by reading and executing the program incorporated into the circuitry. In this embodiment, each processor is not limited to being configured as a single circuit; multiple independent circuits may be combined to form a single processor, and its functions may be realized in this way.

[0114] It should be noted that the components of each device illustrated in the above description of the embodiments are functional concepts and do not necessarily have to be physically configured as shown. In other words, the specific forms of distribution and integration of each device are not limited to those illustrated, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads and usage conditions. Furthermore, all or any part of the processing functions performed by each device can be realized by a CPU and a program that is analyzed and executed by the CPU, or by hardware using wired logic.

[0115] Furthermore, the processing method described in the above-mentioned embodiments can be implemented by executing a pre-prepared processing program on a computer such as a personal computer or workstation. This processing program can be distributed via a network such as the Internet. Alternatively, this processing program can be recorded on a computer-readable non-temporary recording medium such as a hard disk, flexible disk (FD), CD-ROM, MO, DVD, USB memory, and SD card memory, and executed by being read from the non-temporary recording medium by a computer.

[0116] According to at least one embodiment described above, the number of tumor point clusters used for clustering can be optimized, and the efficiency of clustering can be significantly improved while ensuring coverage.

[0117] While several embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0118] 10: Tumor Point Cluster Optimization System 1: Acquisition device 2: Internal sampling device 3: Surface sampling device 4: Output device

Claims

1. A method for optimizing tumor point clouds in tumor ablation planning, which obtains an optimized tumor point cloud for use in tumor ablation planning by optimizing the tumor point cloud, which represents a collection of three-dimensional voxels of a tumor. Acquisition step of acquiring the tumor point group, and dividing the tumor point group into internal points representing a set of three-dimensional voxels inside the tumor and surface points representing a set of three-dimensional voxels on the tumor surface, An internal sampling step is performed to downsample the aforementioned internal points and obtain internal sampling points, A surface sampling step is performed on the surface points to obtain surface sampling points, by performing sampling different from that in the internal sampling step. Output step of outputting the internal sampling points and the surface sampling points as the optimized tumor point group. A method for optimizing tumor point clouds in tumor ablation planning, including the above.

2. The surface sampling step is, A rasterization step comprising rasterizing the surface points using a grid composed of multiple blocks to obtain a surface block which is a block corresponding to each point within the surface points, wherein the size of the block corresponds to a point in the tumor point group, For each grid, a centroid calculation step is performed to calculate the centroid position of each surface block in that grid, The step includes: calculating the distance of each surface block from the centroid position for each grid, and selecting a predetermined number of surface blocks as sampling blocks based on the distance and a weight related to the distance; The tumor point cloud optimization method according to claim 1, wherein in the output step, the sampling block is the surface sampling point.

3. The grid is composed of S * S blocks, In the rasterization step, the minimum 3D bounding box of the tumor point cluster is determined, the surface of the minimum 3D bounding box is rasterized on the grid, and the block containing the tumor surface points in each grid is defined as the surface block. In the sampling block selection step, the Manhattan distance is calculated as the distance, and in the weighted selection, if there are multiple surface blocks at the same distance, the more such blocks there are, the greater the selection weight given to the surface blocks at the same distance. The tumor point group optimization method according to claim 2, wherein the sampling rate in the internal sampling step is inversely proportional to S.

4. The process further includes the step of the user pre-setting one or more designated ablation areas in the tumor point group by performing a preliminary plan, The tumor point group optimization method according to claim 1, wherein in the acquisition step, tumor point groups outside the one or more designated ablation ranges are acquired from the tumor point group and divided into internal points and surface points.

5. A tumor ablation planning method that performs tumor ablation planning based on a tumor point cloud representing a collection of three-dimensional voxels of a tumor, The steps include optimizing the tumor point group using the tumor point group optimization method described in claim 1 to obtain an optimized tumor point group, A planning step of performing a tumor ablation plan using the optimized tumor point group. Tumor ablation planning methods, including those mentioned above.

6. The tumor ablation planning method according to claim 5, wherein in the planning step, the user specifies in advance the number of ablations and performs the ablation plan.

7. A tumor point cloud optimization system that obtains an optimized tumor point cloud for use in tumor ablation planning by optimizing the tumor point cloud, which represents a collection of three-dimensional voxels of a tumor. An acquisition unit that acquires the tumor point group and divides the tumor point group into internal points representing a collection of three-dimensional voxels inside the tumor and surface points representing a collection of three-dimensional voxels on the tumor surface, An internal sampling unit that downsamples the aforementioned internal points to obtain internal sampling points, A surface sampling unit that performs sampling on the surface points differently from the internal sampling unit to obtain surface sampling points, An output unit that outputs the internal sampling points and the surface sampling points as the optimized tumor point group. A tumor point cluster optimization system equipped with the following features.

8. A tumor ablation planning system that performs tumor ablation planning based on a tumor point cloud representing a collection of three-dimensional voxels of a tumor, A tumor point group optimization system according to claim 7, which optimizes the tumor point group to obtain an optimized tumor point group, A planning unit that uses the optimized tumor point group to perform ablation planning for the tumor. A tumor ablation planning system equipped with the following features.

Images