Double-lumen bronchial catheter entry simulation method, system and device and storage medium

Abstract

The invention provides a double-lumen bronchial catheter entry simulation method, system and device and a storage medium. The method comprises the steps that a chest thin layer CT scanning sequence is converted into mask data; converting the mask data into a bronchus three-dimensional model; extracting a center line model of the bronchial three-dimensional model; constructing a graph data structure; calculating the shortest path from the starting point to the end point of the inlet pipe, and generating a path point list; rendering a temporarily selected double-lumen bronchial catheter on the shortest path; and along with the tube entering operation, comparing the size relationship among the airway radius, the tracheal catheter radius and the bronchial catheter radius in real time, and outputting the tube entering and exiting animation and the tube diameter comparison result. According to the invention, the specification optimization of the double-lumen bronchial catheter can be realized in a simulated catheter entering manner, and the safety and effectiveness of the intubation operation can be improved in an auxiliary manner; according to the method and the device, the intubation process is visually simulated before intubation, so that a visual reference is provided for path planning, and the intubation failure risk caused by an improper insertion position can be reduced.

Description

Technical Field

[0001] The present application relates to the field of medical equipment technology, and specifically to a double-lumen bronchial tube insertion simulation method, system, equipment and storage medium. Background Art

[0002] A double-lumen endobronchial tube is a medical tube with two separate lumens (e.g. Figure 3 Double-lumen endobronchial tubes (DLTs) are primarily used in thoracic surgery and some cardiac surgeries to achieve lung isolation and unilateral lung ventilation (collapse the operative lung and keep it static to facilitate operation, while maintaining normal ventilation of the non-operative lung). Double-lumen endobronchial tubes significantly improve the safety and controllability of thoracic surgery and reduce intraoperative complications. They are an indispensable tool in modern anesthesiology and thoracic surgery. However, the failure rate of double-lumen endobronchial tube insertion using traditional techniques ranges from 32% to 83%.

[0003] In actual clinical application, the key to successful double-lumen endobronchial tube intubation is: on the one hand, the specifications of the double-lumen endobronchial tube must match the patient's airway anatomy. If the tube diameter is too thin, the airway resistance will increase, the non-operative lung ventilation will be poor, and intraoperative hypoxia may occur easily; if the tube diameter is too thick, the risk of airway damage during intubation will increase; on the other hand, the insertion position of the double-lumen endobronchial tube must be very accurate (such as Figure 4 If the insertion position is improper, ideal single-lung ventilation cannot be achieved, and the lung on the surgical side expands, which in turn leads to difficult surgical operation or intraoperative hypoxia.

[0004] Currently, the selection of double-lumen endobronchial tube specifications and the accurate insertion of the target position are highly dependent on the anesthesiologist's operational experience and preoperative imaging evaluation. However, some anesthesiologists lack knowledge of imaging, which makes it difficult to select double-lumen tube specifications and predict the intubation position, thereby affecting the intubation success rate and surgical safety.

[0005] Therefore, how to design a double-lumen bronchial tube insertion simulation system that can avoid the mismatch between the specifications of the double-lumen bronchial tube and the anatomical structure of the patient's airway and improve the accuracy of the intubation position is a technical problem that has not yet been solved in the existing technology. Summary of the Invention

[0006] Therefore, the technical problem to be solved by this application is to overcome the technical defects in the prior art in that the selection of double-lumen bronchial tube specifications and the prediction of intubation position are highly dependent on the experience and imaging evaluation of anesthesiologists, resulting in improper specification matching and poor intubation position accuracy, thereby providing a double-lumen bronchial tube insertion simulation method, system, equipment and storage medium that can avoid the mismatch between the specification selection of the double-lumen bronchial tube and the anatomical structure of the patient's airway and improve the accuracy of the intubation position.

[0007] This application mainly includes the following aspects: In a first aspect, an embodiment of the present application provides a double-lumen endobronchial catheter insertion simulation method, the method comprising: acquiring a chest thin-layer CT scan sequence; converting the chest thin-layer CT scan sequence into mask data; converting the mask data into a bronchial three-dimensional model through a moving cube algorithm; extracting a centerline model of the bronchial three-dimensional model through a mean curvature flow algorithm; constructing a graph data structure based on the bronchial three-dimensional model and the centerline model, and making the centerline nodes of the centerline model correspond one-to-one to the vertices of the graph data structure, the vertices including the spatial coordinates of the centerline nodes and the surface points of the bronchial three-dimensional model before contraction corresponding to the centerline nodes. Coordinate list; obtain the spatial coordinates of the entry start point and the end point of the centerline model, calculate the shortest path from the entry start point to the end point through a graph search algorithm, and generate a path point list; obtain the specification parameters of the temporarily selected double-lumen bronchial tube, and render the temporarily selected double-lumen bronchial tube on the shortest path; follow the entry operation of the temporarily selected double-lumen bronchial tube, update the front end coordinates of the tracheal tube and the front end coordinates of the bronchial tube in real time, calculate the airway radius corresponding to the front end coordinates of the tracheal tube and the front end coordinates of the bronchial tube in real time, compare the size relationship between the airway radius and the tracheal tube radius and the bronchial tube radius, and output the entry animation and tube diameter comparison result.

[0008] According to one embodiment of the present application, the step of calculating in real time the airway radius corresponding to the coordinates of the front end of the tracheal tube and the coordinates of the front end of the bronchial tube, respectively, and comparing the size relationship between the airway radius and the tracheal tube radius and the bronchial tube radius includes: obtaining the first coordinate value of the next coordinate in the path point list corresponding to the coordinates of the front end of the bronchial tube; calculating the first airway radius of the airway position corresponding to the first coordinate value; comparing the first airway radius with the bronchial tube radius, if the first airway radius is smaller than the bronchial tube radius, displaying the tube diameter comparison result in red; obtaining the second coordinate value of the next coordinate in the path point list corresponding to the coordinates of the front end of the tracheal tube; calculating the second airway radius of the airway position corresponding to the second coordinate value; comparing the second airway radius with the tracheal tube radius, if the second airway radius is smaller than the tracheal tube radius, displaying the tube diameter comparison result in red.

[0009] According to one embodiment of the present application, the step of calculating the first airway radius of the airway position corresponding to the first coordinate value includes: extracting the first spatial coordinate corresponding to the first coordinate value and the first surface point coordinate list corresponding to the first spatial coordinate from the graph data structure; calculating the Euclidean distance between the first spatial coordinate and all surface point coordinates in the first surface point coordinate list, and taking the minimum value as the first airway radius; the step of calculating the second airway radius of the airway position corresponding to the second coordinate value includes: extracting the second spatial coordinate corresponding to the second coordinate value and the second surface point coordinate list corresponding to the second spatial coordinate from the graph data structure; calculating the Euclidean distance between the second spatial coordinate and all surface point coordinates in the second surface point coordinate list, and taking the minimum value as the second airway radius.

[0010] According to one embodiment of the present application, the chest thin-layer CT scan sequence is converted into the mask data through a deep learning image segmentation model; the deep learning image segmentation model is a three-dimensional segmentation neural network constructed based on the 3DU-Net network; the training parameters of the three-dimensional segmentation neural network include: a sample batch size of 2, a voxel block size of , the normalization scheme adopts CTNormalization, the median voxel size is [234,509.5,512], and the image space spacing is [1,0.7333984375,0.7333984375]; the network parameters of the three-dimensional segmentation neural network include: the number of basic feature channels is 32, the number of convolution layers of each stage of the encoder is [2,2,2,2,2,2], and the number of convolution layers of each stage of the decoder is [2,2,2 ,2,2,2], the number of pooling times for each axis is [4,5,5], the downsampling step is 2, the convolution kernel size is [[3,3,3],[3,3,3],[3,3,3],[3,3,3],[3,3,3],[3,3,3]], the pooling kernel size is [[1,1,1],[2,2,2],[2,2,2],[2,2,2],[1,2,2]], the convolution step is 1, and the convolution padding is 1.

[0011] According to one embodiment of the present application, the steps of obtaining the specification parameters of the temporarily selected double-lumen bronchial catheter and rendering the temporarily selected double-lumen bronchial catheter on the shortest path include: establishing an index value list for the shortest path; obtaining the specification parameters of the temporarily selected double-lumen bronchial catheter and determining the front end position of the tracheal catheter; rendering the double-lumen bronchial catheter on the shortest path based on the index value corresponding to the front end of the double-lumen bronchial catheter on the shortest path, and after the front end position of the tracheal catheter passes the entry starting point, rendering the tracheal catheter model and the bronchial catheter model separately with the front end position of the tracheal catheter as the dividing point.

[0012] According to one embodiment of the present application, after rendering the tracheal tube model and the bronchial tube model respectively with the front end position of the tracheal tube as the dividing point, the step of rendering the bronchial tube model at the tracheal carina smoothly is also included, specifically including: if the current bronchial tube front end index value is greater than the tracheal carina point index value, it indicates that the bronchial tube part has entered the bronchus; in the constructed vtkPoints instance, the index point of the tracheal carina point index value and several index points before and after it are deleted; the vtkPoints instance after deleting the index point is input as a control point into vtkParametricSpline to generate a parameterized curve equation; the parameterized curve equation is input into vtkParametricFunctionSource for interpolation to obtain the interpolated discrete curve data; a vtkTubeFilter instance is constructed, the interpolated discrete curve data is input, the radius is set to the bronchial tube radius, and the end sealing parameter is set to True; vtkTubeFilter The output is used as the final rendering data and sent to the VTK rendering pipeline for rendering to achieve a smooth transition rendering effect from the main trachea to the bronchi.

[0013] According to one embodiment of the present application, the specification parameters of the double-lumen endobronchial catheter include: endobronchial catheter length; endotracheal catheter radius; endotracheal catheter radius.

[0014] In a second aspect, the embodiment of the present application further provides a double-lumen bronchial tube insertion simulation system, the system comprising: an acquisition module for acquiring a chest thin-layer CT scan sequence, the spatial coordinates of the tube insertion start and end points, and the specification parameters of a temporarily selected double-lumen bronchial tube; an image segmentation module for converting the chest thin-layer CT scan sequence into mask data; a three-dimensional modeling module for converting the mask data into a three-dimensional bronchial model; a centerline extraction module for extracting a centerline model from the three-dimensional bronchial model; a graph data construction module for constructing a graph data structure; and a path planning module. Module: used to calculate the shortest path from the entry start point to the end point and generate a list of path points; model rendering module: used to render the tracheal tube model and the bronchial tube model, including: a smooth transition processing unit: used to achieve a smooth transition of the bronchial tube model at the tracheal carina position; a radius calculation and comparison module: used to calculate in real time the airway radius corresponding to the front end coordinates of the tracheal tube and the front end coordinates of the bronchial tube, and compare the size relationship between the airway radius and the tracheal tube radius and the bronchial tube radius; a display module: used to display the entry animation and tube diameter comparison results.

[0015] In a third aspect, an embodiment of the present application further provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the one processor, and the instructions are executed by the at least one processor so that the steps of the above-mentioned method are implemented when the at least one processor executes the instructions.

[0016] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-described method are implemented.

[0017] The technical solution provided in this application has the following advantages: 1. The double-lumen endobronchial tube insertion simulation method of the present application is based on converting a chest thin-layer CT scan sequence into a three-dimensional bronchial model and a centerline model. After obtaining the specification parameters of the temporarily selected double-lumen endobronchial tube, the insertion is simulated, and the insertion operation is followed, and the relationship between the airway radius and the tracheal tube radius and the endobronchial tube radius is continuously compared step by step, thereby achieving the optimization of the specifications of the double-lumen endobronchial tube; the physician directly judges the compatibility of the temporarily selected double-lumen endobronchial tube with the patient's airway based on the comparison results, which can effectively avoid the risk of airway damage caused by too coarse specifications or ventilation problems caused by too fine specifications, and help improve the safety and effectiveness of intubation operations.

[0018] 2. The double-lumen endobronchial tube insertion simulation method of the present application, by integrating medical image processing technology and three-dimensional modeling technology, performs a visual simulation of the double-lumen endobronchial tube insertion process before intubation, providing an intuitive visual reference for planning the double-lumen endobronchial tube insertion path, helping to reduce the risk of intubation failure due to improper insertion position and reducing dependence on the experience of the anesthesiologist.

[0019] 3. Compared with the traditional method that requires calculating the normal direction through principal component analysis and searching for the nearest point on the surface point by point, the double-lumen endobronchial tube insertion simulation method of the present application pre-constructs a graph data structure and performs Euclidean distance calculation based on the list of surface point coordinates corresponding to the centerline nodes, thereby converting the complex radius solution into a fast retrieval of fixed data and simple geometric operations, significantly simplifying the real-time calculation process of the airway radius, improving the response efficiency of the simulation system, and providing efficient and accurate technical support for the dynamic evaluation of the tube diameter adaptability during intubation.

[0020] 4. The double-lumen bronchial tube insertion simulation method of the present application renders the tracheal part and the bronchial part of the double-lumen bronchial tube as independent units, avoiding the joint calculation of complex geometric structures in the traditional overall rendering mode, significantly reducing the computing load of the rendering engine, improving the rendering speed, and ensuring the smoothness of the simulation process.

[0021] 5. The double-lumen bronchial tube insertion simulation method of the present application deletes the index points near the carina and reconstructs the parameterized curve, so that the direction of the bronchial tube model at the tracheal carina fits the physiological curvature characteristics of the human airway, avoids image distortion due to geometric mutations, and improves simulation accuracy.

[0022] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 This is the overall flow chart of the double-lumen endobronchial tube insertion simulation method of the present application.

[0025] Figure 2 This is the system architecture diagram of this application.

[0026] Figure 3 It is a physical schematic diagram of an existing double-lumen endobronchial tube.

[0027] Figure 4 This is a schematic diagram of the double-lumen endobronchial tube being inserted into the appropriate position.

[0028] Figure 5 It is a schematic diagram of the binary mask data.

[0029] Figure 6 is a schematic diagram of a voxel cube unit.

[0030] Figure 7 This is a schematic diagram of the 3D rendering interface after selecting the starting and ending points of the entry tube.

[0031] Figure 8 Schematic diagram of the tracheal tube model and the bronchial tube model.

[0032] Figure 9 is a schematic diagram showing the results of pipe diameter comparison. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of illustration and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts can be implemented out of sequence, and steps without logical context can be reversed or implemented simultaneously. In addition, those skilled in the art, under the guidance of the contents of this application, can add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0034] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0035] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. Example

[0036] like Figure 1 As shown, this embodiment provides a double-lumen endobronchial tube insertion simulation method, comprising the following steps: S101: Acquire a thin-slice chest CT scan sequence; S102: Converting a chest thin-slice CT scan sequence into mask data; S103: converting the mask data into a 3D bronchial model using a marching cube algorithm; S104: extracting the centerline model of the bronchial 3D model using the mean curvature flow algorithm; S105: Constructing a graph data structure based on the bronchial 3D model and the centerline model, and making a one-to-one correspondence between centerline nodes of the centerline model and vertices of the graph data structure, wherein the vertices include the spatial coordinates of the centerline nodes and a list of surface point coordinates of the bronchial 3D model before contraction corresponding to the centerline nodes; S106: Obtain the spatial coordinates of the entry start point and the end point of the centerline model, calculate the shortest path from the entry start point to the end point using a graph search algorithm, and generate a path point list; S107: Obtaining specification parameters of a temporarily selected double-lumen endobronchial tube, and rendering the temporarily selected double-lumen endobronchial tube on the shortest path; S108: Following the insertion operation of the temporarily selected double-lumen endobronchial tube, the tracheal tube front end coordinates and the endobronchial tube front end coordinates are updated in real time, the airway radius corresponding to the tracheal tube front end coordinates and the endobronchial tube front end coordinates are calculated in real time, and the size relationship between the airway radius and the tracheal tube radius and the endobronchial tube radius is compared, and the insertion animation and tube diameter comparison results are output.

[0037] The following describes in detail the steps of the double-lumen endobronchial tube insertion simulation method: S101: Acquire a thin-slice chest CT scan sequence; Here, the chest thin-slice CT scan sequence refers to the tomographic data stored in the nii.gz format obtained by performing a thin-slice scan (slice thickness is usually ≤1mm) of the patient's chest using clinical CT equipment. These data retain metadata such as the voxel coordinates and spatial spacing of the original image, and can accurately present the subtle anatomical features of airway structures such as the main trachea and bronchi. S102: Converting a chest thin-slice CT scan sequence into mask data; In one possible implementation, step S102 converts the chest thin-slice CT scan sequence into mask data using a deep learning image segmentation model; the deep learning image segmentation model is a three-dimensional segmentation neural network constructed based on a 3D U-Net network; The training parameters of the 3D segmentation neural network include: The batch size is 2 and the patch size is , the normalization scheme adopts CT Normalization, the median image size in voxels is [234, 509.5, 512], and the image space spacing is [1, 0.7333984375, 0.7333984375]; The network parameters of the 3D segmentation neural network include: The number of basic feature channels (Unet_base_num_features) is 32, the number of convolution layers in each stage of the encoder (n_conv_per_stage_encoder) is [2,2,2,2,2,2], the number of convolution layers in each stage of the decoder (n_conv_per_stage_decoder) is [2,2,2,2,2,2], the number of pooling times per axis (num_pool_per_axis) is [4,5,5], and the downsampling step size (Downsampling The stride is 2, the convolution kernel size (conv_kernel_sizes) is [[3,3,3],[3,3,3],[3,3,3],[3,3,3],[3,3,3],[3,3,3]], the pooling kernel size (pool_op_kernel_size) is [[1,1,1],[2,2,2],[2,2,2],[2,2,2],[1,2,2]], the convolution step (Stride convolution) is 1, and the convolution padding (Padding) is 1.

[0038] This step uses a deep learning image segmentation model to accurately identify the airway area from the chest thin-slice CT scan sequence and convert it into binary mask data (i.e., a matrix with the target area as 1 and the background as 0, such as Figure 5 This mask data can intuitively mark the distribution of the airways in three-dimensional space, providing clear contour boundaries for the subsequent construction of a three-dimensional bronchial model using the marching cube algorithm, thus avoiding the subjectivity and errors of manual segmentation.

[0039] It should be noted that this deep learning image segmentation model uses a three-dimensional segmentation neural network built based on the 3D U-Net network. The 3D U-Net network is a classic architecture in the field of medical image segmentation. Through the "encoder-decoder" structure combined with jump connections, it can not only extract deep features of the image but also retain shallow spatial information.

[0040] S103: converting the mask data into a 3D bronchial model using a marching cube algorithm; Specifically, first input the mask data generated in step S102, and divide the mask data into multiple The voxel cube unit (such as Figure 6As shown in the figure, each unit contains 8 voxels (values ​​are 0 or 1). For each cubic unit, if its 8 voxels contain both 0 and 1 (i.e., there is an airway boundary), the intersection of the voxel edge and the airway surface is calculated (the intersection coordinates are determined by linear interpolation). Then, based on the distribution of 0 and 1 voxels in the unit, a predefined configuration table is queried to determine the connection method of the intersection, generate triangular patches, and fit the airway surface contour. Finally, all cubic units are traversed and the generated triangular patches are spliced ​​into a complete bronchial 3D model.

[0041] Compared with traditional manual modeling, the marching cubes algorithm can automatically generate a complete bronchial tree model from mask data, avoiding manual intervention and improving modeling efficiency. In addition, the algorithm can capture the subtle structure of the airway through voxel-level intersection calculations, ensuring the consistency of the three-dimensional model with the patient's actual anatomical structure, and providing a reliable basis for matching the specifications of double-lumen tubes.

[0042] S104: extracting the centerline model of the bronchial 3D model using the mean curvature flow algorithm; Here, the 3D bronchial model (surface triangular mesh) generated in step S103 is first input. The mean curvature flow algorithm is used to shrink the 3D model. The model surface is iteratively evolved, causing surface points to move along its mean curvature, ultimately converging to the model's centerline skeleton. This process can be implemented using the extract_mean_curvature_flow_skeleton function in the CGAL library. Specifically, during the iterative process, the surface mesh in the intermediate shrinkage stage is called the mid-skeleton. After each iteration, the system locally re-meshes the mid-skeleton through angle segmentation and edge shrinkage. The process terminates when the change in the mid-skeleton between iterations is sufficiently small. Each time an edge is shrunk, two vertices are merged into a single vertex. During this process, the algorithm records the triangular vertices of the original surface mesh and assigns them to the centerline nodes on the final fused centerline.

[0043] The mean curvature flow algorithm can accurately preserve the branch point positions and branch angles of the bronchial tree, avoid the centerline bifurcation point offset caused by algorithm errors, and provide a true anatomical reference for the construction of the image data structure in the subsequent step S105.

[0044] S105: Constructing a graph data structure based on the bronchial 3D model and the centerline model, and making a one-to-one correspondence between centerline nodes of the centerline model and vertices of the graph data structure, wherein the vertices include the spatial coordinates of the centerline nodes and a list of surface point coordinates of the bronchial 3D model before contraction corresponding to the centerline nodes; Here, we first input the 3D bronchial model generated in step S103 and the centerline model extracted in step S104. We then construct a weighted graph data structure using the Boost library's adjacency matrix data structure, namely boost::adjacency_list. Specifically, we use each centerline node in the centerline model as a vertex in the graph data structure. By traversing the centerline topology, we establish a one-to-one correspondence between vertices and centerline nodes.

[0045] Each vertex stores the following attributes: 1) Spatial coordinates: records the position of the centerline node in three-dimensional space, which is used to locate airway anatomical structures (such as carina and bronchial branch points); 2) Surface point coordinate list: stores the coordinates of the airway surface points corresponding to the centerline node in the original three-dimensional model before contraction; it should be emphasized that this list is retained during contraction through the mean curvature flow algorithm, and the airway surface geometry information can be obtained without additional calculation.

[0046] In addition, the edges of the graph data structure are generated according to the connection relationship of the centerline nodes. The weight of the edge is defined as the Euclidean distance between two adjacent centerline nodes, which is used to reflect the actual length of the airway.

[0047] S106: Obtain the spatial coordinates of the entry start point and the end point of the centerline model, calculate the shortest path from the entry start point to the end point using a graph search algorithm, and generate a path point list; The spatial coordinates of the starting and ending points of the tube entry are obtained by clicking the bronchial centerline model with the mouse in the 3D rendering interface (e.g. Figure 7 However, since the location clicked by the user may deviate from the centerline model, when the user clicks with the mouse, the system traverses all vertices of the graph data structure, calculates the three-dimensional Euclidean distance between the click location and the vertex, and selects the point with the smallest distance as the entry start or end point.

[0048] Among them, the graph search algorithm adopts the breadth-first search algorithm (such as breadth_first_search of the boost library) and uses the edges and weights of the graph data structure to calculate the shortest path.

[0049] The calculation of the shortest path does not rely on manual experience, but is directly generated by the algorithm based on the patient's individualized airway model, which improves the objectivity of path planning. Combined with the optimized implementation of the boost library, path calculation can be completed in milliseconds, meeting the needs of clinical real-time simulation.

[0050] S107: Obtaining specification parameters of a temporarily selected double-lumen endobronchial tube, and rendering the temporarily selected double-lumen endobronchial tube on the shortest path; In a possible implementation, the specification parameters of the double-lumen endobronchial tube in step S107 include: bronchial tube length (bronc_len), tracheal tube radius (trachea_rad), and bronchus tube radius (bronchus_rad).

[0051] It should be noted that the length of the main trachea in adults is approximately 11-13 cm (from the glottis to the tracheal carina), while the tracheal tube portion of the double-lumen endobronchial tube commonly used in clinical practice is usually 23-29 cm long. This length design has reserved sufficient margin to ensure that the tracheal tube can be inserted from the mouth and cover the entire main trachea, without the need to consider the issue of insufficient length.

[0052] In a possible implementation, the step of obtaining the specification parameters of the tentatively selected double-lumen endobronchial tube and rendering the tentatively selected double-lumen endobronchial tube on the shortest path in step S107 includes the following steps: Step S1071: Create an index value list for the shortest path; Step S1072: Obtaining the specification parameters of the temporarily selected double-lumen endobronchial tube and determining the front end position of the endobronchial tube; Step S1073: Based on the index value corresponding to the front end of the double-lumen endobronchial tube on the shortest path, the double-lumen endobronchial tube is rendered on the shortest path. After the front end of the endobronchial tube passes the starting point of the tube entry, the endobronchial tube model and the endobronchial tube model are rendered separately with the front end of the endobronchial tube as the dividing point.

[0053] Here, the shortest path point list is defined as path_vec; the index value of the front end of the double-lumen endobronchial tube is idx_front; the index value of the front end of the tracheal tube is idx_bifur; the bronchial tube length is bronc_len; the tracheal tube radius is trachea_rad; and the bronchial tube radius is bronchus_rad.

[0054] The index value list established in step S1071 is an ordered number mapping of the shortest path point list path_vec generated in S106. Specifically, the index value list is an integer sequence starting from 0, [0, 1, 2, ..., n-1], where n is the total number of vertices in the shortest path point list path_vec. Each index value i corresponds to a unique spatial coordinate of path_vec[i], forming a one-to-one mapping relationship of "index-vertex". For example, in clinical scenarios, the oropharynx (airway entrance) is often selected as the starting point, and the target bronchus (such as the first-order branch of the left main bronchus) is selected as the end point. Therefore, the index value of the oropharynx is 0, and the index value of the first-order branch of the left main bronchus is n-1. If the endotracheal tube does not enter the oropharynx (starting point), its index value is negative, and the endotracheal tube model is not rendered.

[0055] like Figure 8 As shown, during rendering, the tracheal tube model is rendered between index 0 and idx_bifur, and the bronchial tube model is rendered between idx_bifur and idx_front. When rendering the tracheal tube, the centerline between 0 and idx_bifur is first found. Then, using this centerline as a reference, the model is moved left and right by half the radius of the tracheal tube, perpendicular to the centerline, to generate two auxiliary centerlines, path_left and path_right. The left and right sections of the tracheal tube are then rendered using these two auxiliary centerlines as axes, and finally merged into the entire tracheal tube model. When rendering the bronchial tube, if the front end of the bronchial tube does not reach the tracheal carina, it is rendered as an extension of one of the auxiliary centerlines. If the front end of the bronchial tube reaches the tracheal carina, steps S10731 through S10736 are executed to achieve a smooth transition from the main trachea to the bronchi.

[0056] This embodiment renders the tracheal and bronchial portions of a double-lumen endobronchial tube as independent units, avoiding the joint calculation of complex geometric structures in the traditional overall rendering mode. This significantly reduces the computational load of the rendering engine, improves the rendering speed, and ensures the smoothness of the simulation process.

[0057] In one possible implementation, after rendering the tracheal tube model and the bronchial tube model separately with the front end of the tracheal tube as a dividing point in step S1073, the step of rendering the bronchial tube model in a smooth transition at the tracheal carina is further included. Specifically, the steps include: Step S10731: If the current endobronchial tube front end index value is greater than the tracheal carina index value, it indicates that the endobronchial tube has partially entered the bronchus; Step S10732: In the constructed vtkPoints instance, delete the index point of the tracheal carina point index value and several index points before and after it; Step S10733: input the vtkPoints instance after deleting the index point as a control point into vtkParametricSpline to generate a parameterized curve equation; Step S10734: Input the parameterized curve equation into vtkParametricFunctionSource for interpolation to obtain the interpolated discrete curve data; Step S10735: Construct a vtkTubeFilter instance, input the interpolated discrete curve data, i.e., the output of vtkParametricFunctionSource, set the radius to the bronchial tube radius, and set the end cap parameter to True; Step S10736: The output of vtkTubeFilter is used as the final rendering data and sent to the vtk rendering pipeline for rendering to achieve a smooth transition rendering effect from the main trachea to the bronchi.

[0058] In this embodiment, by deleting index points near the carina and reconstructing the parameterized curve, the direction of the bronchial tube model at the tracheal carina conforms to the physiological curvature characteristics of the human airway, avoiding image distortion caused by geometric mutations and improving simulation accuracy.

[0059] S108: Following the insertion operation of the temporarily selected double-lumen endobronchial tube, the tracheal tube front end coordinates and the endobronchial tube front end coordinates are updated in real time, the airway radius corresponding to the tracheal tube front end coordinates and the endobronchial tube front end coordinates are calculated in real time, and the size relationship between the airway radius and the tracheal tube radius and the endobronchial tube radius is compared, and the insertion animation and tube diameter comparison results are output.

[0060] It should be noted that the double-lumen endobronchial tube insertion operation is performed through mouse interaction. Specifically, pressing the right mouse button and sliding the mouse to the right in the 3D rendering window will cause the double-lumen endobronchial tube model to move forward one unit along the airway centerline, that is, idx_front will increase by 1. Conversely, pressing the right mouse button and sliding the mouse to the left will cause the double-lumen endobronchial tube model to move back one unit along the airway centerline, that is, idx_front will decrease by 1.

[0061] When the model reaches the end point, continuing to slide to the right will no longer update the index value of idx_front and remain locked. At this time, the front end of the double-lumen endobronchial tube model is fixed to the end point coordinate of path_vec to avoid rendering anomalies caused by out-of-bounds indexes.

[0062] In one possible implementation, the step of calculating in real time the airway radius corresponding to the coordinates of the front end of the tracheal tube and the coordinates of the front end of the bronchial tube, and comparing the size relationship between the airway radius and the radius of the tracheal tube and the radius of the bronchial tube in step S108 includes the following: Case 1: If the endobronchial tube reaches the entry point, but the endotracheal tube does not reach the entry point, the steps include: Step 1081a: Obtain the first coordinate value of the next coordinate in the path point list corresponding to the coordinate of the front end of the bronchial tube; Step 1082a: Calculate a first airway radius of the airway position corresponding to the first coordinate value; Step 1083a: Compare the first airway radius with the bronchial tube radius. If the first airway radius is smaller than the bronchial tube radius, display the tube diameter comparison result in red. Case 2: If the endobronchial tube has passed the entry point and the endotracheal tube has reached the entry point, the steps include: Step 1081b: Obtain the first coordinate value of the next coordinate in the path point list corresponding to the coordinate of the front end of the bronchial tube; Step 1082b: Calculate a first airway radius of the airway position corresponding to the first coordinate value; Step 1083b: Compare the first airway radius with the bronchial tube radius. If the first airway radius is smaller than the bronchial tube radius, display the tube diameter comparison result in red. Step 1084b: Obtain the second coordinate value of the next coordinate in the path point list corresponding to the coordinate of the front end of the tracheal tube; Step 1085b: Calculate the second airway radius of the airway position corresponding to the second coordinate value; Step 1086b: Compare the second airway radius with the tracheal tube radius. If the second airway radius is smaller than the tracheal tube radius, display the tube diameter comparison result in red.

[0063] like Figure 9 As shown in the figure, when the doctor finds that the tube diameter comparison result is displayed in red, it means that the radius of the current double-lumen bronchial tube exceeds the inner diameter of the corresponding airway position, and there is a risk of intubation injury; after that, the doctor should select the next model of double-lumen bronchial tube and enter the new specification parameters. The system will automatically rebuild the double-lumen bronchial tube model and re-perform the tube diameter comparison; it should be noted that the double-lumen bronchial tube finally selected should be the largest specification that is smaller than and closest to the patient's airway inner diameter, rather than blindly selecting the thinnest model with the smallest tube diameter, because if the tube diameter is too thin, it may lead to increased airway resistance, poor lung ventilation on the non-surgical side, and easy hypoxia during surgery.

[0064] In one possible implementation, the steps of calculating the first airway radius of the airway position corresponding to the first coordinate value in steps 1082a and 1082b include the following steps: Step 10821: extracting a first spatial coordinate corresponding to the first coordinate value and a first surface point coordinate list corresponding to the first spatial coordinate from the graph data structure; Step 10822: Calculate the Euclidean distance between the first spatial coordinate and all surface point coordinates in the first surface point coordinate list, and take the minimum value as the first airway radius; In one possible implementation, the step of calculating the second airway radius of the airway position corresponding to the second coordinate value in step 1085b includes the following steps: Step 10851: extracting the second spatial coordinate corresponding to the second coordinate value and the second surface point coordinate list corresponding to the second spatial coordinate from the graph data structure; Step 10852: Calculate the Euclidean distance between the second spatial coordinate and all surface point coordinates in the second surface point coordinate list, and take the minimum value as the second airway radius.

[0065] It should be noted that the traditional method for calculating the airway radius is to first determine the point to be measured on the center line. , and then As the center point, obtain several nearby points before and after ( , , , , ...), and then through principal component analysis, we get Vector of the tangent direction of the point , rotate this vector ,get The vector of the point normal direction , and find the nearest point on the surface of the original bronchial model in the normal direction , find the Euclidean distance between the two points, that is, get the airway radius of the current point.

[0066] The double-lumen endobronchial tube model requires calculating the airway radius at each point along the airway centerline. However, traditional calculation methods, from neighboring point sampling to principal component analysis, require coordinate centering, covariance matrix calculation, and eigendecomposition. For example, using 10 neighboring points, solving the covariance matrix requires a 10×3 matrix multiplication. Eigendecomposition involves iterative operations, which are computationally expensive and require high hardware computing power.

[0067] Compared with the traditional method that requires calculating the normal direction through principal component analysis and searching for the nearest point on the surface point by point, this application pre-constructs a graph data structure and performs Euclidean distance calculation based on the list of surface point coordinates corresponding to the centerline nodes, thereby converting the complex radius solution into a fast retrieval of fixed data and simple geometric operations. This significantly simplifies the real-time calculation process of the airway radius, improves the response efficiency of the simulation system, and provides efficient and accurate technical support for the dynamic evaluation of the tube diameter adaptability during intubation.

[0068] like Figure 2As shown, based on the same application concept, the embodiment of the present application also provides a double-lumen bronchial tube insertion simulation system, including: an acquisition module: used to obtain a chest thin-layer CT scan sequence, the spatial coordinates of the insertion start and end points, and the specifications of a temporarily selected double-lumen bronchial tube; an image segmentation module: used to convert the chest thin-layer CT scan sequence into mask data; a three-dimensional modeling module: used to convert the mask data into a three-dimensional bronchial model; a centerline extraction module: used to extract a centerline model from the three-dimensional bronchial model; a graph data construction module: used to construct a graph data structure; a path planning module: used to calculate the shortest path from the insertion start to the end point and generate a list of path points; a model rendering module: used to render the tracheal tube model and the bronchial tube model, including: a smooth transition processing unit: used to achieve a smooth transition of the bronchial tube model at the tracheal carina position; a radius calculation and comparison module: used to calculate in real time the airway radius corresponding to the tracheal tube front end coordinates and the bronchial tube front end coordinates, respectively, and compare the size relationship between the airway radius and the tracheal tube radius and the bronchial tube radius; and a display module: used to display the insertion animation and tube diameter comparison results.

[0069] Based on the same application concept, an embodiment of the present application also provides a computer device, including at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by a processor, and the instructions are executed by the at least one processor so that the above method is implemented when the at least one processor executes the instructions.

[0070] Based on the same application concept, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored, and the above method is implemented when the computer program is executed by a processor.

[0071] In the embodiment of the present application, the computer program can also execute other machine-readable instructions when run by the processor to execute the methods described in other embodiments. For the specific execution method steps and principles, please refer to the description of the embodiment and will not be repeated here.

[0072] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0073] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0074] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0075] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0076] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0077] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed in the present application, or make equivalent replacements for some of the technical features thereof. However, these modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A double-lumen endobronchial tube insertion simulation method, characterized in that: include: Obtain a thin-slice chest CT scan sequence; Converting the chest thin-slice CT scan sequence into mask data; Converting the mask data into a three-dimensional bronchial model by a marching cube algorithm; extracting a centerline model of the three-dimensional bronchial model by using a mean curvature flow algorithm; constructing a graph data structure based on the bronchial three-dimensional model and the centerline model, and making a one-to-one correspondence between centerline nodes of the centerline model and vertices of the graph data structure, wherein the vertices include the spatial coordinates of the centerline nodes and a list of surface point coordinates of the bronchial three-dimensional model before contraction corresponding to the centerline nodes; Obtaining the spatial coordinates of the entry point and the end point of the centerline model, calculating the shortest path from the entry point to the end point using a graph search algorithm, and generating a list of path points; Obtaining specification parameters of a temporarily selected double-lumen endobronchial tube, and rendering the temporarily selected double-lumen endobronchial tube on the shortest path; Following the insertion operation of the temporarily selected double-lumen endobronchial tube, the endotracheal tube front end coordinates and the endobronchial tube front end coordinates are updated in real time, the airway radius corresponding to the endotracheal tube front end coordinates and the endobronchial tube front end coordinates are calculated in real time, and the size relationship between the airway radius and the endotracheal tube radius and the endobronchial tube radius is compared, and the tube insertion animation and tube diameter comparison results are output.

2. The double-lumen endobronchial tube insertion simulation method according to claim 1, characterized in that: The step of calculating in real time the airway radius corresponding to the coordinates of the front end of the tracheal tube and the coordinates of the front end of the bronchial tube, and comparing the size relationship between the airway radius and the radius of the tracheal tube and the radius of the bronchial tube comprises: Obtaining a first coordinate value of a next coordinate in a path point list corresponding to the coordinate of the front end of the bronchial tube; Calculating a first airway radius of the airway position corresponding to the first coordinate value; comparing the first airway radius with the bronchial tube radius, and displaying the tube diameter comparison result in red if the first airway radius is smaller than the bronchial tube radius; and / or, obtaining a second coordinate value of the next coordinate in the path point list corresponding to the coordinate of the front end of the tracheal tube; Calculating a second airway radius of the airway position corresponding to the second coordinate value; The second airway radius is compared with the tracheal tube radius. If the second airway radius is smaller than the tracheal tube radius, the tube diameter comparison result is displayed in red.

3. The double-lumen endobronchial tube insertion simulation method according to claim 2, characterized in that: The step of calculating a first airway radius of the airway position corresponding to the first coordinate value includes: extracting from the graph data structure a first spatial coordinate corresponding to the first coordinate value and a first surface point coordinate list corresponding to the first spatial coordinate; Calculating the Euclidean distance between the first spatial coordinate and all surface point coordinates in the first surface point coordinate list, and taking the minimum value as the first airway radius; And / or, the step of calculating the second airway radius of the airway position corresponding to the second coordinate value includes: extracting from the graph data structure a second spatial coordinate corresponding to the second coordinate value and a second surface point coordinate list corresponding to the second spatial coordinate; Calculate the Euclidean distance between the second spatial coordinate and all surface point coordinates in the second surface point coordinate list, and take the minimum value as the second airway radius.

4. The double-lumen endobronchial tube insertion simulation method according to claim 1, characterized in that: The chest thin-layer CT scan sequence is converted into the mask data through a deep learning image segmentation model; the deep learning image segmentation model is a three-dimensional segmentation neural network constructed based on a 3D U-Net network; The training parameters of the three-dimensional segmentation neural network include: The batch size is 2 and the patch size is , the normalization schemes adopt CT Normalization, the median image size invoxels is [234,509.5,512], and the image space spacing is [1,0.7333984375,0.7333984375]; The network parameters of the three-dimensional segmentation neural network include: The number of basic feature channels (Unet_base_num_features) is 32, the number of convolution layers in each stage of the encoder (n_conv_per_stage_encoder) is [2,2,2,2,2,2], the number of convolution layers in each stage of the decoder (n_conv_per_stage_decoder) is [2,2,2,2,2,2], the number of pooling times per axis (num_pool_per_axis) is [4,5,5], and the downsampling step size (Downsampling The stride is 2, the convolution kernel size (conv_kernel_sizes) is [[3,3,3],[3,3,3],[3,3,3],[3,3,3],[3,3,3],[3,3,3]], the pooling kernel size (pool_op_kernel_size) is [[1,1,1],[2,2,2],[2,2,2],[2,2,2],[1,2,2]], the convolution step (Stride convolution) is 1, and the convolution padding (Padding) is 1.

5. The double-lumen endobronchial tube insertion simulation method according to claim 1, characterized in that: The step of obtaining the specification parameters of the temporarily selected double-lumen endobronchial tube and rendering the temporarily selected double-lumen endobronchial tube on the shortest path includes: Creating an index value list for the shortest path; Obtain the specifications of the tentatively selected double-lumen endobronchial tube and determine the front end position of the endotracheal tube; Based on the index value corresponding to the front end of the double-lumen endobronchial tube on the shortest path, the double-lumen endobronchial tube is rendered on the shortest path. After the front end of the endobronchial tube passes the entry starting point, the endobronchial tube model and the endobronchial tube model are rendered separately with the front end of the endobronchial tube as the dividing point.

6. The double-lumen endobronchial tube insertion simulation method according to claim 5, characterized in that: After rendering the tracheal tube model and the bronchial tube model separately with the front end of the tracheal tube as the dividing point, the step of rendering the bronchial tube model in a smooth transition at the tracheal carina is further included, specifically comprising: If the current endobronchial tube front end index value is greater than the tracheal carina index value, it indicates that the endobronchial tube has partially entered the bronchus; In the constructed vtkPoints instance, delete the index point of the tracheal carina point index value and several index points before and after it; Input the vtkPoints instance after deleting the index point as the control point into vtkParametricSpline to generate the parametric curve equation; Input the parameterized curve equation into vtkParametricFunctionSource for interpolation to obtain the interpolated discrete curve data; Construct a vtkTubeFilter instance, input the interpolated discrete curve data, set the radius to the radius of the bronchial tube, and set the end cap parameter to True; The output of vtkTubeFilter is used as the final rendering data and sent to the vtk rendering pipeline for rendering to achieve a smooth transition rendering effect from the main trachea to the bronchi.

7. The double-lumen endobronchial tube insertion simulation method according to claim 1, characterized in that: The specifications of the double-lumen endobronchial tube include: bronchial tube length (bronc_len); Tracheal tube radius (trachea_rad); Bronchus tube radius (bronchus_rad).

8. A double-lumen endobronchial tube insertion simulation system, characterized in that: include: Acquisition module: used to obtain the chest thin-slice CT scan sequence, the spatial coordinates of the tube insertion start and end points, and the specifications of the temporarily selected double-lumen endobronchial tube; Image segmentation module: used for converting the chest thin-slice CT scan sequence into mask data; A three-dimensional modeling module: used for converting the mask data into a three-dimensional model of the bronchus; Centerline extraction module: used for extracting a centerline model from the bronchial three-dimensional model; Graph data construction module: used to build graph data structure; Path planning module: used to calculate the shortest path from the entry point to the end point and generate a list of path points; Model rendering module: used to render the tracheal tube model and the bronchial tube model, including: smooth transition processing unit: used to achieve smooth transition of the bronchial tube model at the tracheal carina position; Radius calculation and comparison module: used to calculate in real time the airway radius corresponding to the coordinates of the front end of the tracheal tube and the coordinates of the front end of the bronchial tube, and compare the size relationship between the airway radius and the radius of the tracheal tube and the radius of the bronchial tube; Display module: used to display the pipe entry animation and pipe diameter comparison results.

9. A computer device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the one processor, and the instructions are executed by the at least one processor so that the at least one processor implements the steps of the method according to any one of claims 1 to 7 when executing the instructions.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images