Cauterization position determination device, cauterization position determination method, and program

Abstract

To provide a cauterization position determination device or the like suitable for determining a candidate of a position to be cauterized from measured data.SOLUTION: A cauterization position determination device 1 determines candidate cauterization positions, which are candidates for positions to be cauterized in a heart. A delay region processing unit 11 calculates a delay region in which the conduction velocity of excitation is delayed. In a convergence region processing unit 13, an excitation direction calculation unit 31 calculates a conduction direction of an excitation conduction pathway through which excitation is conducted in the heart, a width calculation unit 33 calculates the width of the excitation conduction pathway using the conduction direction, and a convergence region calculation unit 35 calculates the convergence region where the excitation conduction pathway converges from the calculated width. A cauterization position determination processing unit 17 determines candidate cauterization positions using the delay region and the convergence region.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an ablation position determining device, an ablation position determining method, and a program, and more particularly to an ablation position determining device that determines candidate ablation positions that are candidates for positions to be ablated in the heart. [Background technology]

[0002] As described in Patent Document 1, currently, arrhythmia surgery is performed by acquiring electrical information from within the heart and displaying it on a three-dimensional map to identify the treatment site, and the surgeon uses this information to imagine the optimal treatment site. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-84824 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in Patent Document 1 merely displays measured data. It has been found that poor treatment outcomes are caused by a combination of factors such as patient-related factors, such as the presence of complicated arrhythmia circuits due to myocardial damage or the presence of multiple types of arrhythmia circuits, and operator-related factors, such as whether the operator is experienced or not.

[0005] Currently, systems are being developed to overcome this by providing more detailed mapping information, but satisfactory improvements in treatment outcomes have yet to be achieved. Therefore, the development of alternative approaches to improve treatment outcomes is necessary.

[0006] Therefore, an object of the present invention is to provide an ablation position determining device and the like that is suitable for determining candidate ablation positions from measured data. [Means for solving the problem]

[0007] A first aspect of the present invention is an ablation position determination device that determines candidate ablation positions, which are candidates for positions to be ablated in the heart, comprising a delay region processing unit, a convergence region processing unit, and an ablation position determination processing unit, wherein the convergence region processing unit comprises an excitation direction calculation unit, a width calculation unit, and a convergence region calculation unit, wherein the delay region processing unit calculates a delay region in which the conduction velocity of excitation is delayed, and in the convergence region processing unit, the excitation direction calculation unit calculates the conduction direction of an excitation conduction pathway through which excitation is conducted in the heart, the width calculation unit calculates the width of the excitation conduction pathway using the conduction direction, and the convergence region calculation unit calculates a convergence region where the excitation conduction pathway converges using the calculated width, and the ablation position determination processing unit determines the candidate ablation positions using the delay region and the convergence region.

[0008] A second aspect of the present invention is the ablation position determination device of the first aspect, wherein in the convergence region processing unit, the excitation direction calculation unit divides sensing data obtained by measuring the heart into a plurality of clusters and calculates a conduction direction for each of the clusters, and the width calculation unit calculates a width for each cluster using the conduction direction.

[0009] A third aspect of the present invention is the ablation position determination device of the first or second aspect, wherein in the convergence region processing unit, the excitation direction calculation unit sets a plurality of representative points and calculates a conduction direction for each of the representative points, and the width calculation unit calculates a width for each of the representative points.

[0010] A fourth aspect of the present invention is an ablation position determination device according to any one of the first to third aspects, comprising an annular conduction path determination processor, which determines an excitation circuit in which a tachycardia cycle exists as an annular conduction path, and which determines the candidate ablation position using the delay region, the convergence region, and the annular conduction path.

[0011] A fifth aspect of the present invention provides an ablation position determination method in an ablation position determination device that determines candidate ablation positions in the heart, the ablation position determination device comprising: a delay region processing unit; a convergence region processing unit; and an ablation position determination processing unit. The convergence region processing unit comprises an excitation direction calculation unit, a width calculation unit, and a convergence region calculation unit. The method includes a region calculation step in which the delay region processing unit and the convergence region processing unit calculate a delay region in which the conduction velocity of excitation is delayed and a convergence region where excitation pathways converge, respectively, and a determination step in which the ablation position determination processing unit determines the candidate ablation positions using the delay region and the convergence region. In the region calculation step, the excitation direction calculation unit calculates the conduction direction of excitation pathways in the heart, the width calculation unit calculates the width of the excitation pathways using the conduction direction, and the convergence region calculation unit calculates the convergence region where the excitation pathways converge using the calculated width.

[0012] A sixth aspect of the present invention is the ablation position determination method of the fifth aspect, wherein in the convergence region processing unit, the excitation direction calculation unit divides sensing data obtained by measuring the heart into a plurality of clusters and calculates a conduction direction for each cluster, and the width calculation unit calculates a width for each cluster using the conduction direction.

[0013] A seventh aspect of the present invention is the ablation position determination method of the fifth or sixth aspect, wherein in the convergence region processing unit, the excitation direction calculation unit sets a plurality of representative points and calculates a conduction direction for each of the representative points, and the width calculation unit calculates a width for each of the representative points.

[0014] An eighth aspect of the present invention is a method for determining an ablation position according to any one of the first to third aspects, wherein the ablation position determination device includes an annular conduction path determination processor that determines an excitation circuit in which a tachycardia cycle exists as an annular conduction path, and the ablation position determination processor determines the candidate ablation position using the delay region, the convergence region, and the annular conduction path.

[0015] A ninth aspect of the present invention is a program for causing a computer to function as the ablation position determination device of any one of the first to fourth aspects. The present invention may also be understood as a computer-readable recording medium on which the program of the ninth aspect is recorded. [Effects of the Invention]

[0016] According to each aspect of the present invention, the ablation position determination device can determine candidate ablation positions using the delay region and the convergence region. Therefore, for example, the device can be incorporated into a 3D mapping system used in catheter ablation (arrhythmia surgery), a radical procedure for atrial flutter, and used as an assist function to guide the surgeon to the optimal treatment site. This is expected to improve treatment outcomes for atrial flutter by improving patient- and surgeon-related factors that worsen treatment outcomes. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a block diagram showing an example of the configuration of an ablation position determining device 1 according to an embodiment of the present invention. [Figure 2] 2 is a flowchart showing an example of the operation of the ablation position determining device 1 of FIG. 1. FIG. [Figure 3] FIG. 3 is a diagram for explaining the processing of step STA2 in FIG. [Figure 4] FIG. 3 is a diagram for explaining calculation of the delay region in FIG. 2(b). [Figure 5] FIG. 2 is a diagram illustrating an example of the position of each point cloud data included in the sensing data. [Figure 6] FIG. 6 is a diagram showing, as a whole, the conduction direction Q1 in which excitation propagates and the region Q2 in which excitation conduction pathways converge in the sensing data of FIG. 5. [Figure 7] FIG. 3 is a diagram for explaining the calculation of width by dividing the sensing data in FIG. 2(c) by density clustering, which is unsupervised learning. [Figure 8]8 is an enlarged view of cluster R1 in FIG. 7, showing the conduction direction S1 calculated for cluster R1. [Figure 9] FIG. 1 is a first diagram for explaining calculation of width from the direction of excitation on the heart in FIG. 2(d). [Figure 10] FIG. 2(d) is a second diagram for explaining calculation of the width from the direction of excitation on the heart. [Figure 11] FIG. 3 is a third diagram for explaining calculation of width from the direction of excitation on the heart in FIG. 2(d). [Figure 12] FIG. 4 is a fourth diagram for explaining calculation of the width from the direction of excitation on the heart in FIG. 2(d). [Figure 13] FIG. 5 is a fifth diagram for explaining calculation of width from the direction of excitation on the heart in FIG. 2(d). [Figure 14] FIG. 10 is a diagram showing a location W1 that was actually cauterized by a doctor. [Figure 15] 10A and 10B are diagrams for explaining density clustering-based and vector-based ablation position estimation results. [Figure 16] FIG. 2 is a block diagram showing an example of the configuration of an ablation position determining device 51 which is another example of an embodiment of the present invention. [Figure 17] 17 is a flowchart showing an example of the operation of the ablation position determining device 51 of FIG. 16. FIG. [Figure 18] 10 is a diagram for explaining how a width calculation unit 61 calculates a width. FIG. [Figure 19] FIG. 10 is a diagram illustrating a process for calculating a propagation direction vector of excitation on a mesh. [Figure 20] FIG. 10 is a diagram illustrating a process of calculating a streamline from a vector field using an arbitrary initial value on a mesh. [Figure 21] 10A and 10B are diagrams illustrating a process of detecting cycles by depth-first search in a graph obtained by connecting streamlines. [Figure 22] FIG. 10 shows the most highly predicted position (top 1) when the calculation of direction vectors and widths in this embodiment is adopted and the annular conduction path is not taken into account. [Figure 23]10 is a diagram showing a circular conduction path obtained when the calculation of the direction vector and width in this embodiment is adopted and the circular conduction path is taken into consideration. FIG. [Figure 24] FIG. 10 is a diagram showing the most highly predicted position (top 1) when the calculation of direction vectors and widths in this embodiment is adopted and a ring-shaped conduction path is taken into account. DETAILED DESCRIPTION OF THE INVENTION

[0018] Hereinafter, embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited to these embodiments. [Example]

[0019] FIG. 1 is a block diagram showing an example of the configuration of an ablation position determining device 1 according to an embodiment of the present invention.

[0020] Surgery for arrhythmia involves difficult cases where the treatment site varies from case to case. The optimal treatment site for this disease is often a site where intracardiac excitation propagation is delayed and a site where excitation pathways converge. Here, excitation pathways are the paths through which excitation stimuli travel to cause the heart to move, such as beating. The ablation position determination device 1 detects both the site where intracardiac excitation propagation is delayed and the site where excitation pathways converge, and combines the results to estimate candidate ablation sites.

[0021] The ablation position determining device 1 includes a processing device 3, a storage device 5, and an input / output device 7.

[0022] The processing device 3 is a processing device that operates under the control of a program such as a central processing unit (CPU). The processing device 3 includes a delay domain processing unit 11, a convergence domain processing unit 13, a control processing unit 15, and an ablation position determination processing unit 17. Each processing unit can be realized by the processing device operating under the control of a program.

[0023] The delay region processing unit 11 includes a distance calculation unit 21 , a time phase difference calculation unit 23 , a point-to-point velocity calculation unit 25 , and a delay region calculation unit 27 .

[0024] The convergence region processing unit 13 includes an excitation direction calculation unit 31, a width calculation unit 33, and a convergence region calculation unit .

[0025] The storage device 5 is a storage device that stores data, such as a hard disk, a memory, etc. The storage device 5 includes a cardiac shape data storage unit 41, a sensing data storage unit 43, a delay region storage unit 45, a convergence region storage unit 47, and an ablation position storage unit 49.

[0026] The input / output device 7 is a device that performs input / output processing of data with the outside (such as a user or a device other than the cauterization position determining device 1).

[0027] FIG. 2 is a flow chart showing an example of the operation of the ablation position determining device 1 of FIG.

[0028] FIG. 2(a) is a flow chart showing an example of the overall operation of the ablation position determining device 1 of FIG.

[0029] The input / output device 7 stores the cardiac shape data in the cardiac shape data storage unit 41 and stores the sensing data in the sensing data storage unit 43 (step STA1). The cardiac shape data is data indicating the shape of the heart. The sensing data is three-dimensional point cloud data sensed during surgery. Each point cloud data included in the sensing data includes an electric potential, a time phase, and a position.

[0030] The control processing unit 15 performs pre-processing (step STA2) by mapping the sensing data onto cardiac shape data and removing outliers.

[0031] The delay domain processing unit 11 and the convergence domain processing unit 13 calculate the delay domain and the convergence domain (step STA3).

[0032] The delay region is a region where intracardiac excitation propagation is delayed. The delay region processing unit 11 detects the delay region by calculating the conduction velocity based on the distance on the heart and the time phase difference between two point cloud data included in the sensing data.

[0033] The convergence region is the region where excitation pathways converge. In the convergence region processing unit 13, the excitation direction calculation unit 31 and width calculation unit 33 calculate the conduction direction and width, and the convergence region calculation unit 35 calculates the convergence region. Below, we will explain how the convergence region processing unit 13 calculates the convergence region using two examples. One is to divide the sensing data using density clustering and calculate the width. The other is to calculate the width from the excitation direction on the heart.

[0034] The ablation-position determination processor 17 determines ablation candidate positions, which are candidates for positions to be ablated in arrhythmia surgery (catheter ablation), using the delay region and the convergence region (step STA4). The ablation-position determination processor 17 determines the ablation candidate positions, for example, by a region where the delay region and the convergence region are common.

[0035] FIG. 2(b) is a flow diagram showing an example of the operation of the delay domain processing unit 11 of FIG.

[0036] The distance calculation unit 21 and the time phase difference calculation unit 23 calculate the distance and the time phase difference (step STB1).

[0037] Point-to-point velocity calculation unit 25 calculates the conduction velocity (step STB2).

[0038] The delay region calculation unit 27 calculates the delay region (step STB3).

[0039] 2(c) and 2(d) are flow diagrams for explaining an example of the operation of the convergence region processing unit 13 in FIG.

[0040] FIG. 2(c) is a flow chart showing an example of the operation of the convergence region processing unit 13 in FIG.

[0041] The excitation direction calculation unit 31 divides the point cloud data included in the sensing data into a plurality of clusters (step STC1).

[0042] The excitation direction calculation unit 31 calculates the conduction direction (step STC2).

[0043] The width calculation unit 33 calculates the width of the cluster (step STC3).

[0044] The convergence region calculation unit 35 calculates the convergence region (step STC4).

[0045] FIG. 2(d) is a flowchart showing another example of the operation of the convergence region processing unit 13 in FIG.

[0046] The excitation direction calculation unit 31 sets a representative point (step STD1).

[0047] The excitation direction calculation unit 31 calculates the conduction direction of the representative point (step STD2).

[0048] The width calculation unit 33 defines the box (step STD3).

[0049] The width calculation unit 33 performs filtering (step STD4).

[0050] The width calculation unit 33 calculates the width (step STD5).

[0051] The convergence region calculation unit 35 calculates the convergence region (step STD6).

[0052] FIG. 3 is a diagram for explaining the process of step STA2 in FIG. 2(a).

[0053] Fig. 3(a) is a diagram showing the shape of the heart identified by the heart shape data, and Fig. 3(b) is a diagram showing the state after each point cloud data included in the sensing data is mapped onto the heart shape data of Fig. 3(a) and processing such as outlier removal has been performed.

[0054] FIG. 4 is a diagram for explaining the calculation of the delay region in FIG. 2(b).

[0055] The distance calculation unit 21 calculates the Euclidean distance x between two point cloud data using the positions of each point cloud data included in the sensing data. The time difference calculation unit 23 calculates the time difference Δt between each point cloud data using the time phases of each point cloud data included in the sensing data.

[0056] For a value R of a predetermined radius and a value s of a time phase, the inter-point velocity calculation unit 25 calculates the conduction velocity v between each point cloud data (time phase t0) and points within a radius R and with a time phase (t < t0 + s) as v = x / Δt. The delay region calculation unit 27 sets the point cloud data with a slow conduction velocity (for example, the lowest 1%) as the delay region.

[0057] Referring to FIG. 4, the process of calculating the conduction velocity v between the point cloud data P1 and P2 will be described. The distance calculation unit 21 calculates the Euclidean distance x between the point cloud data P1 and P2. Assume the distance is 10. Assume the time phases of the point cloud data P1 and P2 are 60 and 50 respectively. The time difference calculation unit 23 calculates the time difference Δt between the point cloud data P1 and P2 as 60 - 50. The inter-point velocity calculation unit 25 calculates the conduction velocity v as x / Δt. In this example, with 10 / 10, the conduction velocity is 1.

[0058] Referring to FIGS. 5 to 13, the process of calculating the convergence region will be described.

[0059] FIG. 5 shows an example of the positions of each point cloud data included in the sensing data. The conduction direction and width vary depending on the location. FIG. 6 shows the conduction direction Q1 in which excitation propagates as a whole and the region Q2 where the excitation conduction paths converge in the sensing data of FIG. 5.

[0060] Figures 7 and 8 explain how the sensing data in Figure 2(c) is divided by density clustering, an unsupervised learning method, and the width is calculated. Hereinafter, this will be referred to as "density clustering-based." In density clustering-based methods, the sensing data is divided, and the conduction direction and width are detected as being constant within each cluster.

[0061] 7, the excitation direction calculation unit 31 divides the point cloud data included in the sensing data into four clusters R1, R2, R3, and R4 using density clustering and time phases. Specifically, in addition to density clustering using DBSCAN, the excitation direction calculation unit 31 prevents points with a time phase difference greater than a threshold from being included in the same cluster.

[0062] In a typical DBSCAN, if the number of points within a distance ε from the representative point is equal to or greater than the point count reference value MinPts, the points are considered to be in the same cluster. In the density clustering used in this embodiment, even if a point is within a distance ε from the representative point, if the time phase of that point is separated from the time phase of the representative point by more than the time phase reference value, the point is not included in the cluster regardless of the distance. By taking time phase into consideration, clustering that takes into account the flow of excitation in addition to spatial density becomes possible.

[0063] 8 is an enlarged view of cluster R1 in FIG. 7, showing the conduction direction S1 calculated for cluster R1. For each point cloud data included in cluster R1, the excitation direction calculation unit 31 calculates the conduction direction S1 by the sum of vectors starting from the position of the cluster itself and ending at the point cloud data of the time phase t0+1 following the time phase t0 of the cluster itself.

[0064] In FIG. 8, the height S4 indicates the height of the cluster R1. Among the point cloud data included in the cluster R1, there is point cloud data with the farthest distance. These two pieces of point cloud data that form a pair are called the farthest point pair. The width calculation unit 33 sets, as the height of the cluster, the one with the largest inner product between the vector indicating the conduction direction S1 and the vector of the farthest point pair. In FIG. 8, the width calculation unit 33 calculates the height S4 based on the inner product between the vector indicating the conduction direction S1 and the vectors of the farthest point pair S2 and S3.

[0065] The width calculation unit 33 calculates the area of the heart-shaped data of the cluster R1 (the area of the mesh data of the region where the point cloud within the cluster is distributed). The width calculation unit 33 calculates the width of the cluster R1 by dividing the area of the cluster R1 by the height S4 of the cluster R1.

[0066] The convergence region calculation unit 35 extracts, as the cluster indicating the convergence region, one or more clusters from the side with a smaller width.

[0067] FIGS. 9 to 13 are for explaining the calculation of the width from the direction of excitation on the heart in FIG. 2(d). Hereinafter, it is referred to as "vector base". In the vector base, the width of the excitation conduction path is calculated by obtaining the distribution of the point cloud in the direction perpendicular to the excitation conduction path.

[0068] Referring to FIG. 9, for the phase reference values k and d, the excitation direction calculation unit 31 obtains the point cloud F that satisfies the condition (t0 + k - d < t < t0 + k + d) within the surrounding radius r of each point cloud data (phase t0) included in the sensing data. Among the point cloud F, the excitation direction calculation unit 31 sets the direction of the nearest point as the excitation conduction direction vector of that point. When there is no point that satisfies the condition, the excitation direction calculation unit 31 assumes that that point does not have a direction vector. The excitation direction calculation unit 31 repeats this operation for all the point cloud data.

[0069] Referring to FIG. 10, the excitation direction calculation unit 31 downsamples the sensing data to set representative points. Let the point cloud of the representative points be G.

[0070] Referring to FIG. 11, the excitation direction calculation unit 31 sums the vectors of the points within a radius R around each point of G to obtain the direction vectors of each representative point. In FIG. 11, points T1 and T2 are representative points. The vectors starting from points T1 and T2 are the direction vectors of the representative points.

[0071] Referring to FIG. 12, for each point of G, the width calculation unit 33 calculates the normal vector of the heart surface from the heart shape data of the heart surface around the representative point. The width calculation unit 33 defines rectangular parallelepipeds U1 and U2 that are sufficiently large in the cross product direction of the conduction direction vector of the representative point and the normal vector of the heart surface. The width calculation unit 33 selects a group of points (points within the rectangular parallelepiped with a phase close to that of the representative point) that are included in the rectangular parallelepiped and satisfy (t0 + k - d < t < t0 + k + d) with respect to the representative point (phase t0). Note that the atrial surface may also be calculated.

[0072] Referring to FIG. 13, the width calculation unit 33 creates the smallest rectangular parallelepiped that includes the selected group of points. The width calculation unit 33 calculates an approximation of the geodesic distance from each point of the obtained group of points to the farthest point. Here, the geodesic distance is the surface distance on the heart (atrium). The width calculation unit 33 takes the inner product of the geodesic distance from each point to the farthest point and the width direction vector, and sets the maximum value as the widths V1 and V2 of the excitation conduction path.

[0073] The convergence region calculation unit 35 extracts one or more representative points from the side with a smaller width as the representative points indicating the convergence region.

[0074] The ablation positioning processing unit 17, for example, sets the overlapping region of the delay region and the convergence region as the ablation candidate position.

[0075] FIGS. 14 and 15 are diagrams for explaining the ablation position estimation results based on density clustering and vector base.

[0076] Figure 14 shows the location W1 actually cauterized by a doctor. Figure 15(a) shows the convergence region processing unit W2 estimated based on density clustering. Figure 15(b) shows the convergence region processing unit W3 estimated based on vectors. The inventors confirmed that the location W1 in Figure 14 can be estimated by the overlapping region of the delay region and the convergence region processing units W2 and W3.

[0077] The results of the inventors' experiments are explained below. In the experiment, the top five narrowest overlapping regions of the delay region and the convergence region were used as the estimated position. The convergence region was calculated using DBSCAN, which uses only spatial information, density clustering-based, and vector-based methods.

[0078] For 12 cases of reentrant atrial flutter, the accuracy rate of the estimated location was compared, with the location actually cauterized by the doctor being considered the correct answer.

[0079] For comparison, in the case of DBSCAN, which uses only spatial information, the DBSCAN parameters were individually changed to perform segmentation. This is because, when run with fixed parameters, all point clouds may end up in the same cluster depending on the case. The parameters were set to values ​​that would not result in too few clusters.

[0080] The results showed that both density clustering-based and vector-based methods showed improved accuracy compared to DBSCAN, which uses only spatial information. Vector-based methods were the most accurate at 83.3%, followed by density clustering-based methods at 66.7%, which was an improvement over the 42% accuracy of DBSCAN, which uses only spatial information and requires individual parameter adjustment.

[0081] The visualization results show that DBSCAN, which uses only spatial information, does not properly segment the area for width calculation, with some areas being treated as small clusters or noise, while others generate very large clusters, depending on the location on the heart. As a result, even if a very small cluster happens to overlap with the ablation location, it is often judged to be a narrow area, making it impossible to make accurate estimates.

[0082] In contrast, density clustering does not require individual parameter settings, but does not generate extremely large or small clusters, and can determine the estimated position based on the width of the cluster.

[0083] Furthermore, since vector-based methods do not involve clustering, they do not require parameter setting based on empirical rules like DBSCAN, and can perform flexible estimation even when the density of point clouds differs depending on the region.

[0084] Therefore, compared to DBSCAN estimation, which uses only spatial information, density clustering-based and vector-based methods are superior in terms of estimation accuracy and the elimination of the need for individual parameter settings. [Example]

[0085] FIG. 16 is a block diagram showing an example of the configuration of an ablation position determining device 51 which is another example of an embodiment of the present invention.

[0086] 16, components that are the same as those in the ablation position determination device 1 in Fig. 1 are denoted by the same reference numerals. That is, the ablation position determination device 51 includes a processing device 3, a storage device 5, and an input / output device 7. The delay region processing unit 11 and the control processing unit 15 included in the processing device 3, and the cardiac shape data storage unit 41, the sensing data storage unit 43, the delay region storage unit 45, the convergence region storage unit 47, and the ablation position storage unit 49 included in the storage device 5 basically operate in the same manner as those denoted by the same reference numerals in the ablation position determination device 1 in Fig. 1.

[0087] The ablation position determination device 51 in Fig. 16 differs from the ablation position determination device 1 in Fig. 1 in the calculation of the convergence region by the convergence region processing unit 53. That is, the processing device 3 includes the convergence region processing unit 53 and a ring-shaped conduction pathway determination processing unit 63. The convergence region processing unit 53 includes an excitation direction calculation unit 59, a width calculation unit 61, and a convergence region calculation unit 57.

[0088] FIG. 17 is a flowchart showing an example of the operation of the ablation position determining device 51 of FIG.

[0089] Fig. 17(a) is a flow chart showing an example of the overall operation of the ablation position determination device 51 in Fig. 16. It is basically the same as Fig. 2(a), but the candidate ablation positions are determined by referring to the annular conduction pathway in addition to the delay region and the convergence region.

[0090] The input / output device 7 stores the cardiac shape data in the cardiac shape data storage unit 41 and stores the sensing data in the sensing data storage unit 43 (step STE1). The control processing unit 15 performs preprocessing (step STE2). The delay region processing unit 11, the convergence region processing unit 53, and the annular conduction pathway determination processing unit 63 calculate the delay region, the convergence region, and the annular conduction pathway (step STE3). The ablation position determination processing unit 65 determines the ablation candidate position using the delay region, the convergence region, and the annular conduction pathway (step STE4). The ablation position determination processing unit 65 determines the ablation candidate position, for example, by a region where the delay region, the convergence region, and the annular conduction pathway are common.

[0091] Fig. 17(b) is a flow diagram showing an example of the operation of the delay region processing unit 11 of Fig. 16. It is basically the same as Fig. 2(b). The distance calculation unit 21 and the time phase difference calculation unit 23 calculate the distance and the time phase difference (step STF1). The point-to-point velocity calculation unit 25 calculates the conduction velocity (step STF2). The delay region calculation unit 27 calculates the delay region (step STF3).

[0092] FIG. 17(c) is a flowchart showing an example of the operation of the convergence region processing unit 53 in FIG.

[0093] The excitation direction calculation unit 59 sets a representative point (step STG1). The setting of the representative point can be realized by, for example, the same processing as step STD1 in FIG.

[0094] The excitation direction calculation unit 59 calculates the conduction direction of the representative point (step STG2). In this example, the following data processing is performed to obtain a direction vector m iAsk for.

[0095] The coordinates of the representative point are (x i ,y i ,z i ) and the time phase is t i The set of points around the representative point F i A point f included in j The coordinates of (x j ,y j ,z j ) and the time phase is t j Using equation 1, the coordinate difference D between the representative point and the surrounding points is i and the time phase difference w i Ask for.

[0096]

number

[0097] As shown in Equation 2, the gradient vector g i , the vector D i and weight w i and the point set F i Divide by the number of points J in the direction vector m i the gradient vector g i The direction vector m is used as the conduction direction of the representative point. i By using the above, it is possible to prevent the device from reacting sensitively to noise.

[0098]

number

[0099] The width calculation unit 61 defines a box (step STG3) and performs filtering (step STG4). The box definition and filtering can be realized by the same processes as steps STD3 and STD4 in FIG. 2(d), respectively.

[0100] The width calculation unit 61 calculates the width (step STG5). In this example, the width is calculated by the following data processing.

[0101] Points that are spatially and temporally close to the representative point are extracted. Figure 18 is a diagram for explaining how width calculation unit 61 calculates the width. In Figure 18, the points represented by white circles are extracted points, called extracted points. A perpendicular line is drawn from the extracted point to the direction vector, and the intersection is defined as point h. The point obtained by projecting point h onto the heart surface is defined as point h'. Since the representative point and point h' are points on the heart surface, the geodesic distance between the two points on the heart surface is calculated. The extracted points on both the left and right of the direction vector where the geodesic distance is maximum are found, and the sum of these distances is defined as the width of the conduction path.

[0102] The convergence region calculation unit 57 calculates the convergence region (step STG6).

[0103] In step STE3, the ring-shaped pathway determination processor 63 detects an excitation circuit (ring-shaped pathway). The ring-shaped pathway is an excitation circuit in which a tachycardia cycle exists. In this example, the ring-shaped pathway determination processor 63 detects the ring-shaped pathway by the following data processing.

[0104] The loop conduction path determination processor 63 determines loop conduction paths by streamline analysis. First, it calculates the propagation direction vector of excitation on the mesh (see FIG. 19). Next, it calculates streamlines from the vector field using arbitrary initial values ​​on the mesh (see FIG. 20). Next, it detects cycles in the graph obtained by connecting the streamlines using a depth-first search (see FIG. 21).

[0105] The process of calculating the propagation direction vector of excitation on the mesh will be described with reference to FIG. 19. The loop conduction pathway determination processor 63 calculates the direction vector for each time phase data by the same calculation as that explained in Equation 1 and Equation 2. That is, the coordinates of the point corresponding to each time phase data are calculated as (x i ,y i ,z i ) and the time phase is t i The set of points surrounding each point F i A point f included in j The coordinates of (x j ,y j,z j ) and the time phase is t j Using equation 1, the coordinate difference D between the representative point and the surrounding points is i and the time phase difference w i As shown in equation 2, the gradient vector g i , the vector D i and weight w i and the point set F i Divide by the number of points J in the vector m i the gradient vector g i is normalized. If a calculated value is available, it may be used.

[0106] Referring to Figure 20, the process of calculating streamlines from a vector field using any initial value on a mesh will be explained. FPS (Farthest Point Sampling) is performed on the vertices of the mesh, and the resulting point group is used as the initial value. Streamlines are calculated from the initial value using the Runge-Kutta method. The direction vector of the point nearest to the current position is used as the gradient of that point. Points that have been slightly moved are projected onto the mesh and used.

[0107] With reference to Figure 21, we will explain the process of detecting cycles using a depth-first search in a graph obtained by connecting flow lines. Points that make up flow lines that satisfy the condition of "points with small inter-point distances and small time phase differences" are connected. A depth-first search is performed on the obtained graph to detect cycle parts that form loops. The detected cycle parts are considered to be loop conduction paths.

[0108] In this embodiment, either one of the calculation of the direction vector and width and the consideration of the annular conduction path in the calculation of the convergence region may be adopted, or both may be adopted.

[0109] Inventors conducted experiments and found that in Example 1, the accuracy was 0.75 for the top 3 and 0.833 for the top 5. When the calculation of direction vectors and widths in this example was adopted and the annular conduction path was not taken into account, the accuracy was 0.789 for the top 1, 0.947 for the top 3, and 1.00 for the top 5. When the calculation of direction vectors and widths in this example was adopted and the annular conduction path was also taken into account, the accuracy was 0.895 for the top 1, 0.895 for the top 3, and 0.947 for the top 5.

[0110] Figures 22 to 24 show examples in which prediction failed when the calculation of direction vectors and widths in this embodiment was adopted and the annular conduction path was not taken into account, but prediction was successful when the calculation of direction vectors and widths in this embodiment was adopted and the annular conduction path was taken into account.

[0111] Figure 22 shows the most highly predicted locations (top 1) when the direction vector and width calculations in this example are used without considering the annular pathway, and shows that the top 1 locations are different from the correct ablation locations.

[0112] FIG. 23 shows the obtained circular conduction path when the calculation of the direction vector and width in this embodiment is adopted and the circular conduction path is taken into consideration.

[0113] Figure 24 shows the most highly predicted locations (top 1) when considering the circular pathway using the direction vector and width calculations in this example. It shows that the top 1 locations overlap with the correct ablation locations. [Explanation of symbols]

[0114] 1. Cautery positioning device 3 Processing equipment 5 Storage device 7 Input / Output Devices 11 Delay area processing section 13 Convergence area processing section 15 Control processing section 17. Cautery position determination processing unit 21 Distance calculation section 23 Time phase difference calculation section 25 Point-to-point velocity calculation section 27 Delay area calculation unit 31 Excitation direction calculation part 33 Width calculation section 35 Convergence region calculation section 41 Heart shape data storage unit 43 Sensing data storage unit 45 Delay domain memory 47 Convergence region memory section 49 Cautery position memory unit 51 Cautery positioning device 53 Convergence area processing section 57 Convergence region calculation section 59 Excitation direction calculation part 61 Width calculation part 63 Ring conduction path determination processing unit 65 Ablation position determination processing unit

Claims

1. An ablation position determination device that determines candidate ablation positions that are candidates for positions to be ablated in a heart, a delay domain processor, a convergence domain processor, and an ablation position determination processor; the convergence region processing unit includes an excitation direction calculation unit, a width calculation unit, and a convergence region calculation unit; The delay region processing unit calculates a delay region in which the conduction velocity of the excitation is delayed, In the convergence region processing unit, the excitation direction calculation unit calculates a conduction direction of an excitation conduction pathway through which excitation is conducted in the heart; the width calculation unit calculates a width of the excitatory pathway using the conduction direction; the convergence region calculation unit calculates a convergence region where the excitation pathway converges using the calculated width; The ablation position determination processing unit determines the ablation candidate position using the delay region and the convergence region.

2. In the convergence region processing unit, the excitation direction calculation unit divides sensing data obtained by measuring the heart into a plurality of clusters and calculates a conduction direction for each cluster; The ablation position determination device according to claim 1 , wherein the width calculation unit calculates the width of each cluster using the conduction direction.

3. In the convergence region processing unit, the excitation direction calculation unit sets a plurality of representative points and calculates a conduction direction for each representative point; The ablation position determination device according to claim 1 , wherein the width calculation unit calculates a width for each of the representative points.

4. a ring-shaped conduction path determination processing unit; the annular conduction pathway determination processing unit determines an excitation circuit in which a tachycardia cycle exists as an annular conduction pathway; The ablation position determination device according to claim 1 , wherein the ablation position determination processor determines the candidate ablation position using the delay region, the convergence region, and the annular conduction path.

5. 1. An ablation position determination method for an ablation position determination device that determines candidate ablation positions that are candidates for positions to be ablated in a heart, comprising: The ablation position determination device includes a delay region processing unit, a convergence region processing unit, and an ablation position determination processing unit; the convergence region processing unit includes an excitation direction calculation unit, a width calculation unit, and a convergence region calculation unit; a region calculation step in which the delay region processing unit and the convergence region processing unit respectively calculate a delay region in which the conduction velocity of excitation is delayed and a convergence region in which excitation conduction pathways converge; a determining step in which the ablation position determining processing unit determines the ablation candidate position using the delay region and the convergence region; In the area calculation step, the excitation direction calculation unit calculates a conduction direction of an excitation conduction pathway through which excitation is conducted in the heart; the width calculation unit calculates a width of the excitatory pathway using the conduction direction; The ablation position determination method, wherein the convergence region calculation unit calculates a convergence region where the excitation conduction pathways converge using the calculated width.

6. A program for causing a computer to function as the ablation position determining device according to claim 1.

Images