Hydronephrosis minimally invasive puncture and drainage positioning method

Abstract

The present application relates to the field of image navigation, in particular to a method for positioning of minimally invasive puncture and drainage of hydronephrosis, in the present application, three-dimensional image scanning data of a patient's kidney region is acquired and a kidney region voxel grid is established, the skin, ribs, intestine, renal portal blood vessel region, kidney parenchyma and hydronephrosis region are converted into three-dimensional anatomical structure geometry data, and then a hydronephrosis region graph structure is constructed based on the geometry data of the hydronephrosis region, the chamber volume and chamber connectivity of the hydronephrosis chamber are judged, and the optimal drainage target chamber is selected; at the same time, the ribs, intestine and renal portal blood vessel region are taken as puncture avoidance regions, a non-intersection tree node expansion path is generated in the puncturable space, and the puncture total length value, the skin normal direction angle deviation value, the kidney parenchyma internal travel distance and the optimal drainage target chamber surface normal direction angle deviation value of the candidate puncture path are weighted and calculated, thereby realizing quantitative screening of the candidate puncture path.

Description

Technical Field

[0001] This invention relates to the field of image navigation technology, and in particular to a minimally invasive puncture and drainage positioning method for hydronephrosis. Background Technology

[0002] The field of image navigation technology mainly involves medical image acquisition, image recognition, spatial positioning, path planning, and device guidance. Its main purpose is to provide reference information for target site identification, operation path determination, and medical device positioning in the clinical diagnosis and treatment process.

[0003] Among them, the minimally invasive puncture and drainage positioning method for hydronephrosis refers to the method for determining the hydronephrosis area, puncture approach and placement position of drainage device in minimally invasive puncture and drainage operation for patients with hydronephrosis. It usually combines existing technical means such as ultrasound imaging, CT imaging, medical image marking, body surface positioning, puncture path planning and interventional device guidance to provide positioning basis for percutaneous puncture to enter the target hydronephrosis area and place drainage channel.

[0004] Traditional minimally invasive puncture and drainage methods for hydronephrosis, while combining ultrasound imaging, CT imaging, medical image markers, surface localization, and interventional device guidance to provide a basis for percutaneous puncture into the target hydronephrosis area, have several drawbacks. These include irregular hydronephrosis morphology, varying degrees of interconnection between renal pelvis and calyces, and complex spatial relationships between the ribs, intestines, and renal hilar vessels and the puncture path. Furthermore, these methods often rely heavily on imaging observation and operator experience to determine the puncture approach, making it difficult to continuously correlate and assess factors such as chamber volume, chamber connectivity, path length, angular deviation, and the distance traveled within the renal parenchyma. Summary of the Invention

[0005] To achieve the above objectives, the present invention employs the following technical solution: a minimally invasive puncture and drainage positioning method for hydronephrosis, comprising the following steps: S1: Acquire three-dimensional image scan data characterizing the patient's kidney region, divide the kidney region into voxel grids, and determine the three-dimensional anatomical geometric data of the kidney region voxel grids; S2: Based on the three-dimensional anatomical geometric data, select the optimal drainage target chamber located inside the patient's kidney region; S3: Determine the avoidance area in the optimal drainage target chamber, and determine the candidate puncture path from the optimal drainage target chamber; S4: Based on the geometric data of the three-dimensional anatomical structure, a weighted calculation is performed on each of the candidate puncture paths to select the optimal puncture path; S5: Based on the geometric information of the optimal puncture path, construct the positioning parameters for minimally invasive puncture in hydronephrosis.

[0006] As a further aspect of the present invention, the three-dimensional anatomical geometric data includes geometric data of the skin surface, ribs, intestines, renal hilum vessels, renal parenchyma, and hydronephrosis area; the optimal drainage target chamber includes chamber volume and chamber connectivity; the candidate puncture path includes total puncture path length, angle deviation, and travel distance; the optimal puncture path includes geometric information; and the hydronephrosis minimally invasive puncture positioning parameter set includes three-dimensional coordinates of the needle insertion point on the body surface, a three-dimensional vector of the puncture direction, and three-dimensional coordinates of the endpoint within the chamber.

[0007] As a further aspect of the present invention, step S1 specifically comprises: S101: Acquire three-dimensional image scan data of the kidney region during the preoperative diagnosis of the patient, perform spatial registration and voxelization processing on the three-dimensional image scan data, obtain a mesh set containing multiple tissue voxels, and establish a voxel mesh of the kidney region. S102: Based on the preset range of gray values ​​for tissues and organs, determine the affiliation of each tissue voxel in the voxel grid of the kidney region and obtain the tissue voxel affiliation determination result. S103: Based on the determination result of the tissue voxel classification, the tissue voxels belonging to each category are segmented, including the geometric data of the skin, ribs, intestines, renal hilum vascular region, renal parenchyma and hydronephrosis region, to generate three-dimensional anatomical geometric data.

[0008] As a further aspect of the present invention, step S2 specifically comprises: S201: Based on the geometric data of the three-dimensional anatomical structure, all tissue voxels within the geometric data of the hydronephrosis region are regarded as a set of graph theory nodes. Connecting edges are established between two spatially adjacent tissue voxel nodes to construct the graph structure of the hydronephrosis region. The gray value deviation of tissue voxels and the spatial distance between tissue voxels are calculated. The gray value deviation of tissue voxels is divided by the preset maximum gray value deviation to obtain the gray value deviation ratio. The spatial distance between tissue voxels is divided by the preset maximum adjacency distance to obtain the distance ratio. The gray value deviation ratio and the distance ratio are weighted and summed to determine the weight of the node connecting edge. S202: Divide the hydronephrosis region map structure into multiple sub-maps representing independent hydronephrosis chambers, collect the node connection edge weights corresponding to the segmentation boundary of each hydronephrosis chamber sub-map as the total segmentation boundary cost, evaluate the connectivity between the corresponding hydronephrosis chamber and the main renal pelvis region chamber based on the total segmentation boundary cost, and determine the chamber volume of the corresponding hydronephrosis chamber based on the number of tissue voxels in each independent hydronephrosis chamber sub-map; S203: The volume of each fluid-filled cavity and the connectivity between the fluid-filled cavity and the main renal pelvis region are numerically determined, and the fluid-filled cavity with a volume greater than a preset volume value and the highest connectivity between the fluid-filled cavity and the main renal pelvis region is selected as the optimal drainage target cavity.

[0009] As a further aspect of the present invention, step S3 specifically comprises: S301: For the optimal drainage target chamber, call the geometric data of the rib, intestinal tract and renal hilum vessel region in the three-dimensional anatomical structure geometric data as the puncture avoidance area; S302: Randomly select a point from the body surface skin geometry data in the three-dimensional anatomical structure geometry data as the root node, generate random points in the punctureable space and find the tree node with the smallest Euclidean distance to the random point, extend a fixed step length from the target tree node to the random point to generate a new path segment, compare the new path segment with each puncture avoidance area, detect whether spatial intersection occurs, add the endpoint of the new path segment that has not intersected as a new tree node to the tree, and obtain the non-intersecting tree node expansion path; S303: When the end of the extended path branch of a certain non-crossing tree node reaches the internal region of the optimal drainage target cavity, record the complete puncture path, repeat the operation of extending the fixed step size and detecting each cross until the number of generated paths reaches the required set of preset number thresholds, and generate candidate puncture paths.

[0010] As a further aspect of the present invention, step S4 specifically comprises: S401: Extract the total puncture length of each candidate puncture path, the angular deviation between the candidate puncture path and the normal direction of the skin surface in the three-dimensional anatomical geometric data, the distance the candidate puncture path travels in the renal parenchyma, and the angular deviation between the end of the candidate puncture path and the normal direction of the surface of the optimal drainage target chamber, to obtain the multidimensional geometric feature parameters of the candidate path; S402: Based on the multidimensional geometric feature parameters of the candidate path, the total puncture length is divided by a preset length reference value to obtain a first length ratio, the skin surface normal direction angle deviation is divided by a preset maximum angle deviation to obtain a first angle ratio, the renal parenchyma penetration distance is divided by a preset distance reference value to obtain a second length ratio, and the optimal drainage target chamber surface normal direction angle deviation is divided by the maximum angle deviation to obtain a second angle ratio. The first length ratio, the first angle ratio, the second length ratio, and the second angle ratio are weighted and summed according to a preset path weight coefficient to obtain a comprehensive path cost score that characterizes the performance score of the candidate puncture path. S403: Sort all the comprehensive cost scores of the paths by numerical value, and select the candidate puncture path with the lowest comprehensive cost score as the best puncture path.

[0011] As a further aspect of the present invention, step S5 specifically comprises: S501: Extract the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint inside the cavity from the geometric information of the optimal puncture path; S502: The three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint in the cavity are rearranged in the corresponding order to generate a set of minimally invasive puncture positioning parameters for hydronephrosis.

[0012] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by acquiring three-dimensional image scanning data of the patient's kidney region and establishing a voxel mesh of the kidney region, the skin, ribs, intestines, renal hilar vessels, renal parenchyma, and hydronephrosis area are transformed into three-dimensional anatomical geometric data. Then, based on the geometric data of the hydronephrosis area, a hydronephrosis area map structure is constructed. The volume and connectivity of the hydronephrosis chambers are judged to select the optimal drainage target chamber. Simultaneously, by using the ribs, intestines, and renal hilar vessels as puncture avoidance areas, a non-crossing tree node expansion path is generated within the puncture space. The total puncture length, skin surface normal direction angle deviation, renal parenchyma travel distance, and optimal drainage target chamber surface normal direction angle deviation of candidate puncture paths are weighted and calculated. This achieves quantitative screening of candidate puncture paths, determining the optimal puncture path and generating three-dimensional coordinates of the needle insertion point, three-dimensional vector of the puncture direction, and three-dimensional coordinates of the endpoint within the chamber. This provides a repeatable parameterized basis for minimally invasive puncture and drainage positioning in hydronephrosis. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0015] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0016] Please see Figure 1 This invention provides a minimally invasive puncture and drainage positioning method for hydronephrosis, comprising the following steps: S1: Acquire three-dimensional image scan data characterizing the patient's kidney region, divide the kidney region into voxel grids, and determine the three-dimensional anatomical geometric data of the kidney region voxel grids; S2: Based on three-dimensional anatomical geometric data, select the optimal drainage target chamber located inside the patient's kidney region; S3: Determine the avoidance area in the optimal drainage target chamber, and determine the candidate puncture path from the optimal drainage target chamber; S4: Based on the three-dimensional anatomical geometric data, each candidate puncture path is weighted and calculated to select the best puncture path; S5: Construct the positioning parameters for minimally invasive puncture in hydronephrosis based on the geometric information of the optimal puncture path.

[0017] The three-dimensional anatomical geometric data includes the geometric data of the skin, ribs, intestines, renal hilum vessels, renal parenchyma, and hydronephrosis. The optimal drainage target chamber includes the chamber volume and connectivity. The candidate puncture path includes the total length, angle deviation, and travel distance of the puncture path. The best puncture path includes geometric information. The set of minimally invasive puncture positioning parameters for hydronephrosis includes the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint within the chamber.

[0018] Please see Figure 2 Step S1 is as follows: S101: Acquire three-dimensional image scan data of the kidney region during the preoperative diagnosis of the patient, perform spatial registration and voxelization processing on the three-dimensional image scan data, obtain a mesh set containing multiple tissue voxels, and establish a voxel mesh of the kidney region. First, preoperative imaging equipment collects three-dimensional (3D) image scans of the kidney region during the preoperative diagnosis process. This data originates from continuous tomographic imaging examinations already completed during the clinical preoperative diagnosis phase. The data retrieves the grayscale matrix, slice thickness, pixel spacing, scanning direction, patient positioning markers, and examination time. Before input processing, the 3D image scan data is examined layer by layer. If the layers are continuous, the data is read in the original layer order. If missing layers exist, they are registered as unusable, and spatial relationships are established using adjacent complete layers. If a voxel in the 3D image scan data lacks a grayscale value, the grayscale values ​​of adjacent positions on the same layer and the same spatial position on adjacent layers are read and substituted. If an isolated abnormal grayscale value exists at a certain location, the grayscale distribution of continuous tissue voxels surrounding the isolated abnormal grayscale value is read, and smoothing is performed based on the grayscale distribution of the surrounding continuous tissue voxels. During spatial registration, the patient's position markers, slice thickness, pixel spacing, and scanning direction in the 3D image scan data are read, and each image point is mapped to the same patient's 3D coordinate reference. During voxelization, the voxel spatial scale is determined according to the slice thickness and pixel spacing in the 3D image scan data. If the slice thickness and pixel spacing are close, voxels are generated according to the spatial scale corresponding to both slice thickness and pixel spacing. If the slice thickness is greater than the pixel spacing, the inter-slice positions are first mapped to a uniform distance scale, and then the kidney region space is divided into regular voxels. Each tissue voxel records its spatial coordinates, gray value, the slice it belongs to, and the relationship between adjacent tissue voxels, thereby obtaining a mesh set containing multiple tissue voxels and establishing a voxel mesh for the kidney region.

[0019] S102: Based on the preset range of gray values ​​for tissues and organs, determine the affiliation of each tissue voxel in the voxel grid of the kidney region and obtain the tissue voxel affiliation determination result. The preset grayscale value range for organs and tissues is derived from the three-dimensional image scan data of the kidney region during the preoperative diagnosis of the current patient. The operator selects representative continuous voxels from the three-dimensional image scan data for the skin, ribs, intestines, renal hilar vessels, renal parenchyma, and hydronephrosis. After removing boundary mixed voxels, null voxels, and isolated abnormal voxels, the grayscale value distribution of the representative continuous voxels for each organ and tissue is read. The lowest and highest grayscale values ​​that consistently appear in the representative continuous voxels for each organ and tissue are recorded as the corresponding grayscale value range for the organ and tissue. If there are grayscale differences for the same organ and tissue at different levels, the grayscale value distribution of the continuous voxels of the same organ and tissue in the kidney region voxel grid is read, and the grayscale value range of the corresponding organ and tissue is adjusted according to the grayscale value distribution of the continuous voxels. During classification, the gray values ​​of tissue voxels in the kidney region voxel grid are read one by one. If the gray value of a tissue voxel falls within the gray value range of only one tissue organ, the tissue voxel is registered as the corresponding tissue organ category. If the gray value of a tissue voxel falls within the gray value range of two tissue organs, the already classified categories of the adjacent tissue voxels are read, and the number of each classified category in the adjacent tissue voxels is counted. The classified category with the largest number is registered as the tissue voxel's classification category. If the number of each classified category in adjacent tissue voxels is the same, the tissue voxel is read... The grayscale deviation between the grayscale value and the representative continuous voxel grayscale value of each candidate tissue / organ category is used to register the candidate tissue / organ category with the smaller grayscale deviation. If the grayscale value of a tissue / organ does not fall within the grayscale value range of any tissue / organ, the tissue / organ is registered as a tissue / organ to be determined, and the category of the determined tissue / organ category that is closest to the tissue / organ to be determined and has the smallest grayscale difference is read. If the tissue / organ category still cannot be determined, the tissue / organ to be determined is retained and does not proceed to subsequent geometric data generation, and the tissue / organ classification result is finally obtained.

[0020] S103: Based on the tissue voxel classification results, the tissue voxels belonging to each category are segmented, including geometric data of the skin, ribs, intestines, renal hilum vascular region, renal parenchyma and hydronephrosis region, to generate three-dimensional anatomical geometric data. The skin refers to the soft tissue surface of the patient's body that connects to the external space; the ribs refer to the bony structures that the puncture path may encounter; the intestines refer to the digestive tract structures within the abdominal cavity; the renal hilar vascular region refers to the area where the medial renal vessels enter and exit; the renal parenchyma refers to the renal tissue itself; and the hydronephrosis region refers to the low-density space formed by the expansion of fluid within the renal pelvis and calyces. During segmentation, tissue voxels belonging to the skin, ribs, intestines, renal hilar vascular region, renal parenchyma, and hydronephrosis region are collected separately. For each category of tissue voxels, spatial adjacency is examined. If two tissue voxels of the same category are adjacent in three-dimensional coordinates, they are registered in the same continuous region. If a cluster of tissue voxels of the same category is not connected to the main continuous region, the number of tissue voxels in the cluster is counted. The number of tissue voxels in a cluster is calculated by traversing all tissue voxels within the cluster one by one and accumulating them. The method involves counting voxels. The limit for the number of isolated voxels is determined by the spatial resolution of the 3D image scan data and the number of voxels corresponding to the smallest identifiable anatomical segment. If the number of tissue voxels in a cluster of tissue voxels of the same category is greater than the limit for the number of isolated voxels, the cluster is retained. If the number of tissue voxels in a cluster of tissue voxels of the same category is equal to the limit for the number of isolated voxels, the grayscale continuity of the tissue voxels surrounding the cluster is read and checked again. If the number of tissue voxels in a cluster of tissue voxels of the same category is less than the limit for the number of isolated voxels, the cluster is discarded. During boundary generation, the coordinates of boundary tissue voxels adjacent to other categories of tissue voxels or external space are read for each category of tissue voxels. Geometric surfaces are formed according to the spatial adjacency order between boundary tissue voxels, ultimately obtaining the geometric data of the skin, ribs, intestines, renal hilar vessels, renal parenchyma, and hydronephrosis, generating 3D anatomical structure geometric data.

[0021] Please see Figure 3 Step S2 is as follows: S201: Based on the geometric data of the three-dimensional anatomical structure, all tissue voxels in the geometric data of the hydronephrosis area are regarded as a set of graph theory nodes. Connecting edges are established between spatially adjacent tissue voxel nodes to construct the graph structure of the hydronephrosis area. The gray value deviation of tissue voxels and the spatial distance of tissue voxels between adjacent tissue voxel nodes are calculated. The gray value deviation of tissue voxels is divided by the preset maximum gray value deviation to obtain the gray value deviation ratio. The spatial distance of tissue voxels is divided by the preset maximum adjacency distance to obtain the distance ratio. The gray value deviation ratio and the distance ratio are weighted and summed to determine the weight of the node connecting edge. Graph theory nodes refer to the registration points of individual tissue voxels within the geometric data of the hydronephrosis region in the graph structure. Node connections refer to the connection records between two spatially adjacent tissue voxel nodes. The weight of a node connection refers to the connection cost formed by the gray-level difference and spatial distance between adjacent tissue voxel nodes. Spatial adjacency is determined by the three-dimensional coordinates of tissue voxels in the kidney region voxel grid. A node connection is established when two tissue voxel nodes are spatially adjacent; no node connection is established when two tissue voxel nodes are not spatially adjacent. The deviation of tissue voxel grayscale values ​​between adjacent tissue voxel nodes is obtained by comparing the grayscale values ​​of two adjacent tissue voxel nodes. The preset maximum grayscale deviation is derived from the distribution of tissue voxel grayscale values ​​within the geometric data of the current patient's hydronephrosis region. First, the tissue voxel with the lowest and highest grayscale values ​​within the geometric data of the hydronephrosis region is read. Then, the maximum grayscale difference between actual adjacent tissue voxel nodes is read. The value that can cover the grayscale changes of adjacent tissue voxels within the geometric data of the hydronephrosis region is registered as the preset maximum grayscale deviation. If the tissue voxel grayscale deviation is less than the preset maximum grayscale deviation, the tissue voxel grayscale deviation is used in the grayscale deviation ratio calculation. If the tissue voxel grayscale deviation is equal to the preset maximum grayscale deviation, it is registered as the maximum allowable grayscale difference. If the tissue voxel grayscale deviation is greater than the preset maximum grayscale deviation, it is registered according to the upper limit of the preset maximum grayscale deviation. The tissue voxel spatial distance is obtained from the 3D coordinates of two adjacent tissue voxel nodes within the same patient 3D coordinate reference. The preset maximum adjacency distance is determined by the voxel spatial scale determined by voxelization processing and the allowed spatial adjacency methods. If the tissue voxel spatial distance is less than or equal to the preset maximum adjacency distance, the node connection edge is retained; if the tissue voxel spatial distance is greater than the preset maximum adjacency distance, the node connection edge is canceled. The grayscale deviation ratio is obtained from the proportion of the tissue voxel grayscale value deviation relative to the preset maximum grayscale deviation, and the distance ratio is obtained from the proportion of the tissue voxel spatial distance relative to the preset maximum adjacency distance. The weights of the grayscale deviation ratio and the distance ratio are jointly set by the clarity of the hydronephrosis region boundary and the voxel spatial resolution in the 3D image scan data. When the hydronephrosis region boundary is clear, the participation ratio of the grayscale deviation ratio is increased; when the voxel spatial resolution is high, the participation ratio of the distance ratio is increased. The sum of the two participation ratios is 1. The weights of the node connection edges are determined by summing the product of the grayscale deviation ratio and the corresponding participation ratio, and the product of the distance ratio and the corresponding participation ratio.

[0022] S202: Divide the hydronephrosis region map structure into multiple subgraphs representing independent hydronephrosis chambers, collect the weights of the corresponding node connection edges on the segmentation boundary of each hydronephrosis chamber subgraph as the total segmentation boundary cost, evaluate the connectivity between the corresponding hydronephrosis chamber and the main renal pelvis region chambers based on the total segmentation boundary cost, and determine the chamber volume of the corresponding hydronephrosis chamber based on the number of tissue voxels in each independent hydronephrosis chamber subgraph. The hydronephrosis chamber refers to the interconnected or partially interconnected fluid spaces in the geometric data of the hydronephrosis region. The main renal pelvis region chamber refers to the fluid-filled chamber corresponding to the main space of the renal pelvis. The segmentation boundary refers to the set of node connection edges where the connection relationship between two hydronephrosis chambers changes. The total segmentation boundary cost refers to the cumulative value of the node connection edge weights corresponding to the segmentation boundary. During segmentation, starting from the tissue voxel node corresponding to the main renal pelvis region chamber, adjacent tissue voxel nodes are visited along the node connection edges, and the weight of each node connection edge is read. If the weight of the node connection edge is continuously small in the adjacent region, the tissue voxel nodes on both sides of the node connection edge are registered as the internal connection of the same hydronephrosis chamber. If the weight of the node connection edge is relatively large in the local space, the node connection edge is registered as a candidate boundary edge for hydronephrosis chambers. The segmentation boundary is determined based on the weight distribution of all node connection edges in the current patient's hydronephrosis region map structure. First, all node connection edge weights are read and sorted numerically. Then, node connection edges located in high-weight concentrated segments and forming continuous spatial partitions are registered as segmentation boundaries. If the node connection edges in a high-weight concentrated segment form closed or near-closed partitions, an independent hydronephrosis chamber submap is generated. If the node connection edges in a high-weight concentrated segment cannot form continuous partitions, they are retained as the same hydronephrosis chamber submap. The node connection edge weights corresponding to the segmentation boundaries of each hydronephrosis chamber submap are collected, resulting in a total segmentation boundary cost. The smaller the total segmentation boundary cost, the smaller the grayscale difference and spatial distance difference between adjacent tissue voxel nodes at the segmentation boundary, indicating a tighter connection between the corresponding hydronephrosis chamber and the main renal pelvis region chamber. Conversely, the larger the total segmentation boundary cost, the more concentrated the grayscale difference and spatial distance difference between adjacent tissue voxel nodes at the segmentation boundary, indicating a weaker connection between the corresponding hydronephrosis chamber and the main renal pelvis region chamber. The number of tissue voxels in each independent water-filled chamber subgraph is obtained by traversing all tissue voxel nodes in the independent water-filled chamber subgraph one by one and counting them cumulatively. The volume of a single tissue voxel is obtained by the voxel spatial scale determined by voxelization processing. The corresponding volume of the water-filled chamber is determined by calculating the correspondence between the number of tissue voxels and the volume of a single tissue voxel.

[0023] S203: The volume of each hydronephrotic chamber and the connectivity between the hydronephrotic chamber and the main renal pelvis are numerically judged, and the hydronephrotic chamber with a volume greater than the preset volume value and the highest connectivity between the hydronephrotic chamber and the main renal pelvis are selected as the optimal drainage target chamber. The preset volume value is the minimum chamber volume limit for entering the candidate range. The preset volume value is determined by the minimum space the puncture needle tip needs to fall into the hydronephrosis chamber, the spatial scale of the imaging voxels, and the number of continuous tissue voxels that the optimal drainage target chamber can accommodate. The number of continuous tissue voxels that the optimal drainage target chamber can accommodate is obtained by traversing the continuous tissue voxels inside the hydronephrosis chamber and counting the number of tissue voxels that can accommodate the end of the puncture path. If the volume of the hydronephrosis chamber is greater than the preset volume value, the hydronephrosis chamber is registered as a candidate hydronephrosis chamber; if the volume of the hydronephrosis chamber is equal to the preset volume value, the hydronephrosis chamber is registered as a critical candidate hydronephrosis chamber and continues to participate in the comparison of the connectivity between the hydronephrosis chamber and the main renal pelvis region chambers; if the volume of the hydronephrosis chamber is less than the preset volume value, the hydronephrosis chamber is not registered as a candidate hydronephrosis chamber. The connectivity between the hydronephrosis chamber and the main renal pelvis region is represented by the sum of the segmentation boundary costs, with the hydronephrosis chamber having a smaller sum of segmentation boundary costs ranking higher. If the sum of segmentation boundary costs for two hydronephrosis chambers is the same, then the chamber volumes of the two hydronephrosis chambers are compared, with the hydronephrosis chamber with the larger volume ranking higher. If the chamber volumes of the two hydronephrosis chambers are still the same, then the distance from the center point of each hydronephrosis chamber to the punctureable area of ​​the skin surface is read, with the hydronephrosis chamber with the shorter distance ranking higher. The center point of the hydronephrosis chamber is obtained by statistically analyzing the spatial concentration of the three-dimensional coordinates of all tissue voxel nodes within the hydronephrosis chamber, and the punctureable area of ​​the skin surface is obtained by statistically analyzing the area of ​​the skin surface geometry that does not overlap with the puncture avoidance area. After sorting, the hydronephrosis chamber with a volume greater than a preset volume value and the highest connectivity between the hydronephrosis chamber and the main renal pelvis region is selected as the optimal drainage target chamber.

[0024] Please see Figure 4 Step S3 is as follows: S301: For the optimal drainage target chamber, the geometric data of the rib, intestinal tract, and renal hilum vascular region in the three-dimensional anatomical structure geometry data is used as the puncture avoidance area; The puncture avoidance area refers to the anatomical region where the candidate puncture path must not overlap with the spatial occupant. Rib geometry data is used to record the boundary coordinates of bony structures and the voxels of the internal occupant tissue; intestinal geometry data is used to record the boundary coordinates of the abdominal cavity structures and the voxels of the internal occupant tissue; and renal hilar vascular region geometry data is used to record the boundary coordinates of the area where the medial renal vessels enter and exit and the voxels of the internal occupant tissue. During the call, the boundary coordinates and internal occupant tissue voxels of the rib geometry, intestinal geometry, and renal hilar vascular region geometry are read respectively. The coordinate references of the rib, intestinal, and renal hilar vascular region geometry are checked to ensure they match the optimal drainage target chamber. If the coordinate references match, it is directly registered as a puncture avoidance area. If the coordinate references are offset, the coordinate transformation relationship obtained during spatial registration is used to rearrange the coordinates of the rib, intestinal, and renal hilar vascular region geometry. If a boundary in the rib, intestinal, or renal hilar vascular region geometry is locally discontinuous, the coordinates of the adjacent boundaries on both sides of the discontinuous location are read and continuously registered. If the discontinuous location cannot be continuously registered by the coordinates of the adjacent boundaries on both sides, the discontinuous location is registered as an area to be checked and processed as a puncture avoidance area when generating candidate puncture paths.

[0025] S302: Randomly select a point from the surface skin geometry data in the three-dimensional anatomical structure geometry data as the root node. Generate random points in the punctureable space and find the tree node with the smallest Euclidean distance to the random point. Extend a fixed step length from the target tree node towards the random point to generate a new path segment. Compare the new path segment with each puncture avoidance area to detect whether spatial intersection occurs. Add the endpoint of the new path segment without intersection as a new tree node to the tree to obtain the non-intersecting tree node expansion path. If spatial intersection occurs, that is, it is determined that the new path segment overlaps with the rib, intestinal or renal hilar vessel area in any puncture avoidance area in the three-dimensional anatomical space, then perform the removal operation of the new path segment and terminate the growth action of the current branch. Restart the process of generating random points in the punctureable space and search again for the tree node with the smallest Euclidean distance to the random point. Then, perform the action of extending a fixed step length towards the random point again. The root node refers to the starting point of the candidate puncture path on the skin surface. Tree nodes refer to spatial turning points during the expansion of the candidate puncture path. The punctureable space refers to the three-dimensional space between the skin surface geometry and the optimal drainage target chamber, minus the puncture avoidance area. Before random selection, the outer surface coordinates of the skin surface geometry are read, and the Euclidean distance from each outer surface coordinate to the nearest boundary coordinates of the rib, intestinal, and renal hilar vascular regions is read. The preset safety distance is jointly determined by the outer diameter of the puncture needle, the voxel spatial scale of the three-dimensional image scan data, and the outward expansion range of the puncture avoidance area boundary. The outward expansion range of the puncture avoidance area boundary is obtained by extending the boundary coordinates of the rib, intestinal, and renal hilar vascular regions outward at the voxel spatial scale. If the Euclidean distance from the outer surface coordinates of the skin geometry data to the nearest boundary coordinates of any puncture avoidance area is greater than a preset safety distance, then the outer surface coordinates of the skin geometry data are registered as an optional root node. If the Euclidean distance from the outer surface coordinates of the skin geometry data to the nearest boundary coordinates of any puncture avoidance area is equal to the preset safety distance, then the outer surface coordinates of the skin geometry data are registered as a critical root node. If the Euclidean distance from the outer surface coordinates of the skin geometry data to the nearest boundary coordinates of any puncture avoidance area is less than the preset safety distance, then the outer surface coordinates of the skin geometry data are discarded. When generating random points within the punctureable space, the coordinates of the random points originate from the spatial range jointly defined by the outer surface of the skin geometry data, the outer surface of the optimal drainage target cavity, and the boundary of the puncture avoidance area. After generating random points, the three-dimensional coordinates of the existing tree nodes are read, and the Euclidean distance between each random point and the existing tree nodes is calculated. The tree node with the smallest Euclidean distance is determined as the target tree node. When extending a fixed step size from the target tree node towards a random point, the fixed step size is determined by the voxel spatial scale and the thinnest boundary thickness of the puncture avoidance area. The fixed step size is not less than the side length of a single voxel and not greater than the spatial distance corresponding to the thinnest boundary thickness of the puncture avoidance area. If the fixed step size is greater than the spatial distance corresponding to the thinnest boundary thickness of the puncture avoidance area, the fixed step size is adjusted to the maximum step size allowed for detection. If the fixed step size is less than the side length of a single voxel, no new path segment is generated. After a new path segment is generated, sampling points are set on the new path segment according to the voxel space scale. Each sampling point is compared with the voxels of the internal occupant tissue in the rib, intestinal, and renal hilar vascular regions. If none of the sampling points on the new path segment fall into the puncture avoidance area, the endpoint of the new path segment is added to the tree as a new tree node. If any sampling point on the new path segment falls into the puncture avoidance area, it is determined that the new path segment overlaps with the rib, intestinal, or renal hilar vascular region in any puncture avoidance area in the three-dimensional anatomical space. The new path segment is removed, the growth of the current branch is terminated, and random points are regenerated in the puncturable space. If any sampling point on the new path segment is located at the boundary of the puncture avoidance area, it is treated as a spatial intersection. The new path segment is removed, and random points are regenerated in the puncturable space to obtain a non-intersecting tree node expansion path.

[0026] S303: When the end of the extended path branch of a non-crossing tree node reaches the internal region of the optimal drainage target cavity, record the complete puncture path, repeat the fixed step length extension and cross detection operation until the number of generated paths reaches the preset threshold, and generate candidate puncture paths. A complete puncture path refers to a sequence of tree nodes that continuously connect from the root node in the geometric data of the skin surface to the internal region of the optimal drainage target cavity. During arrival determination, the three-dimensional coordinates of the end of the branch of the non-crossing tree node extension path are read and compared with the voxel coordinates of the tissue occupying the internal region of the optimal drainage target cavity. If the three-dimensional coordinates of the end of the branch of the non-crossing tree node extension path are within the range of the voxel of the tissue occupying the internal region of the optimal drainage target cavity, the complete puncture path is recorded. If the three-dimensional coordinates of the end of the branch of the non-crossing tree node extension path are located at the surface boundary of the optimal drainage target cavity, it is registered as a boundary arrival path, and the path continues to extend for a fixed step length in the current direction before another arrival determination is performed. If the three-dimensional coordinates of the end of the branch of the non-crossing tree node extension path are located outside the optimal drainage target cavity, random points are generated within the punctureable space and the path continues to extend for a fixed step length. If the newly extended path segment spatially intersects with the puncture avoidance area, the new path segment is discarded and random points are regenerated. The preset threshold is determined by the number of paths to be compared during the candidate puncture path scoring stage, the number of selectable root nodes in the skin geometry data, and the number of accessible areas on the surface of the optimal drainage target chamber. The number of selectable root nodes in the skin geometry data is obtained by traversing the outer surface coordinates of the skin geometry data and counting the number of coordinates that meet the preset safety distance requirements. The number of accessible areas on the surface of the optimal drainage target chamber is obtained by traversing the boundary coordinates of the optimal drainage target chamber surface and counting the number of boundary coordinates that can be reached by paths extended from non-intersecting tree nodes. If the number of recorded complete puncture paths is less than the preset threshold, the operation of extending by a fixed step size and detecting intersections is repeated. If the number of recorded complete puncture paths is equal to the preset threshold, the generation of new complete puncture paths is stopped. If the number of recorded complete puncture paths exceeds the preset threshold, the complete puncture paths that reach the preset threshold are retained according to the recording order of the first arrival of the complete puncture path in the internal area of ​​the optimal drainage target chamber, and candidate puncture paths are generated.

[0027] Please see Figure 5 Step S4 is as follows: S401: Extract the total puncture length of each candidate puncture path, the angle deviation between the candidate puncture path and the normal direction of the skin surface in the three-dimensional anatomical structure geometric data, the distance the candidate puncture path travels in the renal parenchyma, and the angle deviation between the end of the candidate puncture path and the surface normal direction of the optimal drainage target chamber, and obtain the multidimensional geometric feature parameters of the candidate path. The total puncture length refers to the cumulative spatial distance of the candidate puncture path from the entry point on the body surface to the endpoint within the cavity. The normal direction of the body surface skin refers to the vertical direction of the outer surface of the skin geometry at the entry point on the body surface. The angle deviation of the normal direction of the body surface skin refers to the angle difference between the direction of the starting segment of the candidate puncture path and the normal direction of the body surface skin. The distance traversed within the renal parenchyma refers to the continuous distance of the candidate puncture path within the renal parenchyma geometry. The angle deviation of the normal direction of the optimal drainage target cavity surface refers to the angle difference between the direction of the candidate puncture path when it enters the optimal drainage target cavity and the normal direction of the optimal drainage target cavity surface. When extracting the total puncture length, the spatial distance between adjacent nodes is read in the order of candidate puncture path nodes, and the total puncture length is obtained by accumulating the spatial distances between all adjacent nodes. When extracting the angle deviation value of the normal direction of the skin surface, the outer surface coordinates of the skin surface geometry data near the needle insertion point are read. The local skin surface direction is determined based on the outer surface coordinates of the skin surface geometry data near the needle insertion point. Then, the direction of the starting segment of the candidate puncture path is read, and the angle difference between the direction of the starting segment of the candidate puncture path and the normal direction of the skin surface is compared. When extracting the renal parenchyma penetration distance, the voxels are used on the candidate puncture path. Continuous sampling points are set at the spatial scale. Each continuous sampling point is determined to be within the voxel range of the renal parenchyma geometry data. The distance of the continuous sampling segments within the voxel range of the renal parenchyma geometry data is accumulated as the renal parenchyma travel distance. When extracting the angular deviation between the end of the candidate puncture path and the surface normal direction of the optimal drainage target chamber, the surface coordinates of the optimal drainage target chamber adjacent to the end of the candidate puncture path are read. The surface normal direction of the optimal drainage target chamber is determined based on the surface coordinates of the optimal drainage target chamber. Then, the direction in which the end of the candidate puncture path enters the optimal drainage target chamber is read, and the angular difference between the two is compared. If the surface normal direction of the candidate puncture path cannot be obtained, the reading range of the outer surface coordinates of the skin geometry data near the needle insertion point is expanded. If the surface normal direction still cannot be obtained after expanding the reading range, the candidate puncture path is not included in the path comprehensive cost score. If the end of the candidate puncture path is not in the internal region of the optimal drainage target chamber, the candidate puncture path is not included in the path comprehensive cost score.

[0028] S402: Based on the multidimensional geometric feature parameters of the candidate path, the total puncture length is divided by a preset length reference value to obtain a first length ratio; the deviation value of the skin normal direction angle is divided by a preset maximum angle deviation to obtain a first angle ratio; the distance traveled within the renal parenchyma is divided by a preset distance reference value to obtain a second length ratio; and the deviation value of the surface normal direction angle of the optimal drainage target chamber is divided by the maximum angle deviation to obtain a second angle ratio. The first length ratio, first angle ratio, second length ratio, and second angle ratio are weighted and summed according to preset path weight coefficients to obtain a comprehensive path cost score characterizing the performance score of the candidate puncture path. The preset path weight coefficients include global path length weight, skin puncture posture weight, renal parenchymal injury risk weight, and target chamber injection accuracy weight. The global path length weight is multiplied by the first length ratio, the skin puncture posture weight is multiplied by the first angle ratio, the renal parenchymal injury risk weight is multiplied by the second length ratio, and the target chamber injection accuracy weight is multiplied by the second angle ratio. The preset length reference value is determined by the distribution of the total puncture length values ​​in the candidate puncture path set. First, all candidate puncture paths that do not spatially intersect with the puncture avoidance area and reach the internal region of the optimal drainage target cavity are read. Then, the total puncture length values ​​corresponding to all candidate puncture paths are read, and the values ​​that represent the concentrated distribution of reachable path lengths are recorded as the preset length reference value. The preset maximum angle deviation is determined by the distribution of the normal direction of the needle insertion point on the body surface geometry data, the distribution of the normal direction of the accessible area on the surface of the optimal drainage target cavity, and the maximum deviation angle of the candidate puncture path that can enter the optimal drainage target cavity. First, read the angle deviation distribution of candidate puncture paths that can enter the optimal drainage target chamber. Then, register the maximum angle deviation that can still enter the internal area of ​​the optimal drainage target chamber as the preset maximum angle deviation. The preset distance reference value is determined by the thickness of the renal parenchyma geometric data around the optimal drainage target chamber and the distribution of the traversing distance of the candidate puncture path within the renal parenchyma geometric data. First, read the spatial thickness of the renal parenchyma geometric data around the optimal drainage target chamber. Then, read the concentrated distribution of the traversing distance of the candidate puncture path within the renal parenchyma geometric data. Register the value that can represent the reference traversing distance as the preset distance reference value. The preset path weight coefficient consists of the global path length weight, the surface puncture posture weight, the renal parenchymal injury risk weight, and the target chamber injection accuracy weight, with the total of the four weights being 1. When setting the weights, the thickness of the renal parenchymal geometric data, the needle insertion range of the skin geometric data, and the size of the area that can be entered on the surface of the optimal drainage target chamber are read from the current patient's three-dimensional anatomical geometric data. If the traversable range of the renal parenchymal geometric data is narrow, the renal parenchymal injury risk weight is increased; if the needle insertion range of the skin geometric data is narrow, the surface puncture posture weight is increased; if the area that can be entered on the surface of the optimal drainage target chamber is small, the target chamber injection accuracy weight is increased; if the overall length of the candidate puncture paths varies greatly, the global path length weight is increased.If the first length ratio is less than 1, the total puncture length is lower than the preset length reference value; if the first length ratio is equal to 1, the total puncture length reaches the preset length reference value; if the first length ratio is greater than 1, the total puncture length exceeds the preset length reference value. If the first angle ratio is less than 1, the angle deviation of the skin normal direction is lower than the preset maximum angle deviation; if the first angle ratio is equal to 1, the angle deviation of the skin normal direction reaches the preset maximum angle deviation; if the first angle ratio is greater than 1, the angle deviation of the skin normal direction exceeds the preset maximum angle deviation; if the second length ratio is less than 1... If the second length ratio is less than 1, the distance traveled within the renal parenchyma is lower than the preset reference distance; if the second length ratio is equal to 1, the distance traveled within the renal parenchyma reaches the preset reference distance; if the second length ratio is greater than 1, the distance traveled within the renal parenchyma exceeds the preset reference distance. If the second angle ratio is less than 1, the angle deviation of the surface normal direction of the optimal drainage target chamber is lower than the maximum angle deviation; if the second angle ratio is equal to 1, the angle deviation of the surface normal direction of the optimal drainage target chamber reaches the maximum angle deviation; if the second angle ratio is greater than 1, the angle deviation of the surface normal direction of the optimal drainage target chamber exceeds the maximum angle deviation. The global path length weight is multiplied by the first length ratio, the surface puncture posture weight is multiplied by the first angle ratio, the renal parenchyma injury risk weight is multiplied by the second length ratio, and the target chamber incidence accuracy weight is multiplied by the second angle ratio. The four product results are then summed to obtain the comprehensive path cost score.

[0029] S403: Sort all path comprehensive cost scores by numerical value and select the candidate puncture path with the lowest comprehensive cost score as the best puncture path. Before sorting, the path number, comprehensive path cost score, total puncture length, skin surface normal direction angle deviation, renal parenchymal passage distance, and optimal drainage target chamber surface normal direction angle deviation value are read for each candidate puncture path. If a candidate puncture path lacks any scoring parameter, it is not included in the sorting; if all scoring parameters are complete, it is included in the sorting. During sorting, the comprehensive path cost scores are arranged from smallest to largest. If the comprehensive path cost score of one candidate puncture path is less than that of another, the candidate puncture path with the smaller score is ranked higher; if the comprehensive path cost score of one candidate puncture path is greater than that of another, the candidate puncture path with the larger score is ranked lower; if two candidate puncture paths have the same comprehensive path cost score, the renal parenchymal passage distance is compared, and the candidate puncture path with the shorter renal parenchymal passage distance is ranked higher; if the renal parenchymal passage distances of two candidate puncture paths are the same, the renal parenchymal passage distances are compared, and the candidate puncture path with the shorter renal parenchymal passage distance is ranked higher. If the internal traversal distance values ​​are the same, the deviation values ​​of the angles of the optimal drainage target chamber surface normal direction are compared between the two candidate puncture paths. The candidate puncture path with the smaller deviation value is ranked higher. If the deviation values ​​of the angles of the optimal drainage target chamber surface normal direction are the same for both candidate puncture paths, the total puncture length is compared, and the candidate puncture path with the shorter total puncture length is ranked higher. If the total puncture lengths are the same for both candidate puncture paths, the shortest distance from the needle insertion point on the body surface to the puncture avoidance area is compared, and the candidate puncture path with the larger distance is ranked higher. After sorting by numerical value, the candidate puncture path with the lowest overall cost score is selected as the optimal puncture path.

[0030] Please see Figure 6 Step S5 is as follows: S501: Extract the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint in the cavity from the geometric information of the optimal puncture path; The three-dimensional coordinates of the needle insertion point on the body surface refer to the position of the starting node of the optimal puncture path within the patient's three-dimensional coordinate reference. The three-dimensional vector of the puncture direction refers to the spatial direction record from the needle insertion point on the body surface towards the adjacent node inside the optimal puncture path. The three-dimensional coordinates of the endpoint within the cavity refer to the position of the terminal node of the optimal puncture path within the internal region of the optimal drainage target cavity. When extracting the three-dimensional coordinates of the needle insertion point on the body surface, the first path node of the optimal puncture path is read, and it is checked whether the first path node is located on the outer surface of the body surface geometric data. If the first path node is located on the outer surface of the body surface geometric data, the three-dimensional coordinates of the first path node are registered as the three-dimensional coordinates of the needle insertion point on the body surface. If the first path node is located inside or outside the body surface geometric data, the nearest intersection point of the outer surface of the body surface geometric data to the first path node is found along the optimal puncture path, and the three-dimensional coordinates of the intersection point are registered as the three-dimensional coordinates of the needle insertion point on the body surface. If the intersection point of the outer surface of the body surface geometric data cannot be found, no corresponding positioning parameters are generated. When extracting the 3D vector of the puncture direction, the 3D coordinates of the needle insertion point on the body surface and the 3D coordinates of the adjacent path nodes after the needle insertion point are read. The positional changes from the 3D coordinates of the needle insertion point to the 3D coordinates of the adjacent path nodes in the lateral, longitudinal, and depth directions are calculated, and the positional changes are organized into a spatial direction record. If the adjacent path nodes after the needle insertion point on the body surface are missing, the nearest valid path node to the needle insertion point in the optimal puncture path is read. If the nearest valid path node to the needle insertion point on the body surface is still missing, no corresponding positioning parameters are generated. When extracting the 3D coordinates of the intracavitary endpoint, the last path node of the optimal puncture path is read, and the 3D coordinates of the last path node are compared with the voxel coordinates of the tissue occupying the internal region of the optimal drainage target cavity. If the last path node is located in the internal region of the optimal drainage target cavity, the 3D coordinates of the last path node are registered as the 3D coordinates of the intracavitary endpoint. If the last path node is located on the surface boundary of the optimal drainage target cavity, the nearest path node entering the internal region of the optimal drainage target cavity in the direction of the optimal puncture path is read. If the last path node is located outside the optimal drainage target cavity, no corresponding positioning parameters are generated.

[0031] S502: Rearrange the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint in the cavity in the corresponding order to generate a set of positioning parameters for minimally invasive puncture in hydronephrosis. The minimally invasive puncture positioning parameter set for hydronephrosis refers to structured data used to record the geometric positioning information of the optimal puncture path. When rearranging the corresponding order, the first item is fixed as the three-dimensional coordinates of the needle insertion point on the body surface, the second as the three-dimensional vector of the puncture direction, and the third as the three-dimensional coordinates of the endpoint within the cavity. The patient's three-dimensional coordinate reference, spatial unit, patient position orientation marker, and optimal puncture path number are recorded for the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint within the cavity, respectively. If the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, or the three-dimensional coordinates of the endpoint within the cavity lack spatial units, the spatial units determined during the voxelization process are read and added. If the corresponding direction of the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, or the three-dimensional coordinates of the endpoint within the cavity is missing... If the orientation is inconsistent with the patient's position markings, it is rearranged according to the spatially registered patient orientation markings. If the 3D vector of the puncture direction lacks directional processing records, the 3D coordinates of the needle insertion point on the body surface and the 3D coordinates of the adjacent path nodes after the needle insertion point are reread, and the 3D vector of the puncture direction is regenerated. If the 3D coordinates of the needle insertion point on the body surface, the 3D vector of the puncture direction, and the 3D coordinates of the endpoint within the cavity are all complete, they are registered as a valid set of minimally invasive puncture positioning parameters for hydronephrosis. If the 3D coordinates of the needle insertion point on the body surface, the 3D vector of the puncture direction, or the 3D coordinates of the endpoint within the cavity are missing and cannot be filled in, the corresponding set of minimally invasive puncture positioning parameters for hydronephrosis is not output. The above process can be used in the scenario of minimally invasive puncture and drainage positioning for hydronephrosis in the biomedical engineering industry. Based on the 3D image scanning data of the kidney region during the patient's preoperative diagnosis, the process sequentially completes the establishment of a voxel grid in the kidney region, determination of tissue voxel assignment, generation of 3D anatomical structure geometric data, screening of the optimal drainage target cavity, generation of candidate puncture paths, determination of the best puncture path, and generation of the set of minimally invasive puncture positioning parameters for hydronephrosis.

[0032] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A minimally invasive puncture and drainage positioning method for hydronephrosis, characterized in that, Includes the following steps: S1: Acquire three-dimensional image scan data characterizing the patient's kidney region, divide the kidney region into voxel grids, and determine the three-dimensional anatomical geometric data of the kidney region voxel grids; S2: Based on the three-dimensional anatomical geometric data, select the optimal drainage target chamber located inside the patient's kidney region; S3: Determine the avoidance area in the optimal drainage target chamber, and determine the candidate puncture path from the optimal drainage target chamber; S4: Based on the geometric data of the three-dimensional anatomical structure, a weighted calculation is performed on each of the candidate puncture paths to select the optimal puncture path; S5: Based on the geometric information of the optimal puncture path, construct a set of positioning parameters for minimally invasive puncture in hydronephrosis.

2. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 1, characterized in that, The three-dimensional anatomical geometric data includes geometric data of the skin surface, ribs, intestines, renal hilum vessels, renal parenchyma, and hydronephrosis area. The optimal drainage target chamber includes chamber volume and chamber connectivity. The candidate puncture path includes total puncture path length, angle deviation, and travel distance. The optimal puncture path includes geometric information. The set of minimally invasive puncture positioning parameters for hydronephrosis includes three-dimensional coordinates of the needle insertion point on the body surface, a three-dimensional vector of the puncture direction, and three-dimensional coordinates of the endpoint within the chamber.

3. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 1, characterized in that, Step S1 is as follows: S101: Acquire three-dimensional image scan data of the kidney region during the preoperative diagnosis of the patient, perform spatial registration and voxelization processing on the three-dimensional image scan data, obtain a mesh set containing multiple tissue voxels, and establish a voxel mesh of the kidney region. S102: Based on the preset range of gray values ​​for tissues and organs, determine the affiliation of each tissue voxel in the voxel grid of the kidney region and obtain the tissue voxel affiliation determination result. S103: Based on the determination result of the tissue voxel classification, the tissue voxels belonging to each category are segmented, including the geometric data of the skin, ribs, intestines, renal hilum vascular region, renal parenchyma and hydronephrosis region, to generate three-dimensional anatomical geometric data.

4. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 1, characterized in that, Step S2 is as follows: S201: Based on the geometric data of the three-dimensional anatomical structure, all tissue voxels within the geometric data of the hydronephrosis region are regarded as a set of graph theory nodes. Connecting edges are established between two spatially adjacent tissue voxel nodes to construct the graph structure of the hydronephrosis region. The gray value deviation of tissue voxels and the spatial distance between tissue voxels are calculated. The gray value deviation of tissue voxels is divided by the preset maximum gray value deviation to obtain the gray value deviation ratio. The spatial distance between tissue voxels is divided by the preset maximum adjacency distance to obtain the distance ratio. The gray value deviation ratio and the distance ratio are weighted and summed to determine the weight of the node connecting edge. S202: Divide the hydronephrosis region map structure into multiple sub-maps representing independent hydronephrosis chambers, collect the node connection edge weights corresponding to the segmentation boundary of each hydronephrosis chamber sub-map as the total segmentation boundary cost, evaluate the connectivity between the corresponding hydronephrosis chamber and the main renal pelvis region chamber based on the total segmentation boundary cost, and determine the chamber volume of the corresponding hydronephrosis chamber based on the number of tissue voxels in each independent hydronephrosis chamber sub-map; S203: The volume of each fluid-filled cavity and the connectivity between the fluid-filled cavity and the main renal pelvis region are numerically determined, and the fluid-filled cavity with a volume greater than a preset volume value and the highest connectivity between the fluid-filled cavity and the main renal pelvis region is selected as the optimal drainage target cavity.

5. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 1, characterized in that, Step S3 is as follows: S301: For the optimal drainage target chamber, call the geometric data of the rib, intestinal tract and renal hilum vessel region in the three-dimensional anatomical structure geometric data as the puncture avoidance area; S302: Randomly select a point from the body surface skin geometry data in the three-dimensional anatomical structure geometry data as the root node, generate random points in the punctureable space and find the tree node with the smallest Euclidean distance to the random point, extend a fixed step length from the target tree node to the random point to generate a new path segment, compare the new path segment with each puncture avoidance area, detect whether spatial intersection occurs, add the endpoint of the new path segment that has not intersected as a new tree node to the tree, and obtain the non-intersecting tree node expansion path; S303: When the end of the extended path branch of a certain non-crossing tree node reaches the internal region of the optimal drainage target cavity, record the complete puncture path, repeat the operation of extending the fixed step size and detecting each cross until the number of generated paths reaches the required set of preset number thresholds, and generate candidate puncture paths.

6. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 1, characterized in that, Step S4 is as follows: S401: Extract the total puncture length of each candidate puncture path, the angular deviation between the candidate puncture path and the normal direction of the skin surface in the three-dimensional anatomical geometric data, the distance the candidate puncture path travels in the renal parenchyma, and the angular deviation between the end of the candidate puncture path and the normal direction of the surface of the optimal drainage target chamber, to obtain the multidimensional geometric feature parameters of the candidate path; S402: Based on the multidimensional geometric feature parameters of the candidate path, the total puncture length is divided by a preset length reference value to obtain a first length ratio, the skin surface normal direction angle deviation is divided by a preset maximum angle deviation to obtain a first angle ratio, the renal parenchyma penetration distance is divided by a preset distance reference value to obtain a second length ratio, and the optimal drainage target chamber surface normal direction angle deviation is divided by the maximum angle deviation to obtain a second angle ratio. The first length ratio, the first angle ratio, the second length ratio, and the second angle ratio are weighted and summed according to a preset path weight coefficient to obtain a comprehensive path cost score that characterizes the performance score of the candidate puncture path. S403: Sort all the comprehensive cost scores of the paths by numerical value, and select the candidate puncture path with the lowest comprehensive cost score as the best puncture path.

7. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 1, characterized in that, Step S5 is as follows: S501: Extract the three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint inside the cavity from the geometric information of the optimal puncture path; S502: The three-dimensional coordinates of the needle insertion point on the body surface, the three-dimensional vector of the puncture direction, and the three-dimensional coordinates of the endpoint in the cavity are rearranged in the corresponding order to generate a set of minimally invasive puncture positioning parameters for hydronephrosis.

8. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 5, characterized in that, If the spatial intersection occurs, i.e., it is determined that the coordinates of the new path segment overlap with the coordinates of the rib, the intestinal tract, or the renal hilum vessel region in any of the puncture avoidance areas in the three-dimensional anatomical space, then the removal operation of the new path segment is performed and the growth action of the current branch is terminated. The process of generating the random point in the punctureable space is restarted, and the tree node with the smallest Euclidean distance to the random point is searched again. The action of extending the fixed step length towards the random point is performed again.

9. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 6, characterized in that, The preset path weight coefficients include global path length weight, body surface puncture posture weight, renal parenchymal injury risk weight, and target chamber incidence accuracy weight.

10. The method for minimally invasive puncture and drainage positioning of hydronephrosis according to claim 9, characterized in that, The global path length weight is multiplied by the first length ratio, the body surface puncture posture weight is multiplied by the first angle ratio, the renal parenchymal injury risk weight is multiplied by the second length ratio, and the target chamber incidence accuracy weight is multiplied by the second angle ratio.

Images