Preoperative planning method and system for vascular intervention

Abstract

A preoperative planning method and system for vascular intervention. The method comprises: acquiring a vascular lesion image of a patient (100); performing image segmentation on the vascular lesion image to obtain a plurality of two-dimensional images (200), the plurality of two-dimensional images corresponding to a plurality of different tissues or structures; generating lesion three-dimensional data on the basis of the plurality of two-dimensional images (300); generating a vascular interventional surgical plan on the basis of the lesion three-dimensional data and surgical instrument parameters (400), the vascular interventional surgical plan comprising surgical instrument models, a use order and an instrument path; making a lesion simulation model on the basis of the lesion three-dimensional data (500); and performing in-vitro practice or actual surgery on the basis of the vascular interventional surgical plan and the lesion simulation model, and learning and adjusting the vascular interventional surgical plan in real time on the basis of an in-vitro practice result or an actual surgical condition (600). Thus, a surgeon can fully know the condition of a patient's lesion before surgery, thereby improving the accuracy of a vascular interventional surgical plan, and further improving the safety of surgery.

Description

A method and system for preoperative planning of vascular intervention

[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on November 30, 2023, with application number 202311615767.0 and invention name “A Method and System for Preoperative Planning of Vascular Intervention”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present invention relates to the technical field of medical simulation-assisted diagnosis and treatment, and in particular to a pre-operative planning method and system for vascular intervention. Background Art

[0003] In clinical practice, each patient's blood vessels are unique. While most vascular disease patients can be treated directly with interventional surgery, treatment is extremely challenging for patients with difficult lesions, complex lesion structures, or tortuous vascular treatment pathways. If the surgeon rashly performs surgery without sufficient confidence, significant complications may result, leading to high surgical risks and prolonged procedures. Pre-operative simulation of the lesion and subsequent surgical planning can avoid these issues and improve surgical safety.

[0004] Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for pre-operative planning of vascular intervention, which can improve the accuracy of vascular intervention surgery plans and enhance the safety of the surgery.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for preoperative planning of vascular intervention, comprising:

[0008] Obtain images of the patient's vascular lesions;

[0009] Performing image segmentation on the vascular lesion image to obtain a plurality of two-dimensional images; the plurality of two-dimensional images correspond to a plurality of different tissues or structures;

[0010] generating three-dimensional data of lesions based on multiple two-dimensional images;

[0011] Generate a vascular interventional surgery plan based on the three-dimensional lesion data and surgical instrument parameters; the vascular interventional surgery plan includes surgical instrument models, usage sequence, and instrument paths;

[0012] Creating a lesion simulation model based on the three-dimensional lesion data;

[0013] Based on the vascular intervention surgery plan and the lesion simulation model, in vitro practice or actual surgery is performed, and according to the in vitro practice results or the actual surgery situation, the vascular intervention surgery plan is learned and adjusted in real time.

[0014] Optionally, before performing image segmentation on the vascular lesion image, the vascular intervention pre-operative planning method further includes:

[0015] The vascular lesion image is subjected to denoising and contrast enhancement processing in sequence.

[0016] Optionally, a Gaussian filter is used to perform denoising on the vascular lesion image.

[0017] Optionally, performing image segmentation on the vascular lesion image to obtain multiple two-dimensional images specifically includes:

[0018] Segmenting the vascular lesion image using Otsu's algorithm to determine tissue regions or structural regions in the vascular lesion image to obtain a preliminary image set;

[0019] performing region growing processing on the preliminary image set to obtain a grown target image set;

[0020] Canny edge detection is performed on the grown target image set to obtain multiple two-dimensional images.

[0021] Optionally, generating three-dimensional data of the lesion based on the multiple two-dimensional images specifically includes:

[0022] Stacking multiple two-dimensional images to obtain a preliminary three-dimensional image;

[0023] performing voxelization and point cloud reconstruction on the preliminary three-dimensional image in sequence to obtain a preliminary three-dimensional model of the lesion;

[0024] The preliminary three-dimensional lesion model is optimized to obtain final three-dimensional lesion data.

[0025] Optionally, generating a vascular interventional surgery plan based on the three-dimensional lesion data and surgical instrument parameters specifically includes:

[0026] fusing the three-dimensional data of the lesion with surgical instruments of different models respectively, and adjusting the order of using the surgical instruments of different models to obtain a preliminary fused three-dimensional image;

[0027] performing denoising and alignment processing on the preliminary fused three-dimensional image in sequence to obtain a corrected fused three-dimensional image;

[0028] Using a Canny edge detection algorithm to extract edge features of the surgical area in the corrected fused three-dimensional image to obtain a feature-fused three-dimensional image;

[0029] Using an ICP algorithm to align and match the lesion point cloud and the surgical instrument point cloud in the feature-fused three-dimensional image to obtain an aligned fused three-dimensional image;

[0030] Based on the aligned and fused three-dimensional images, geometric calculations and geometric constraints are used to determine the surgical instrument model, usage sequence, and instrument path.

[0031] To achieve the above object, the present invention also provides the following solution:

[0032] A vascular intervention preoperative planning system includes: an image acquisition device, a central processing unit, and a 3D printing terminal; the central processing unit is connected to the image acquisition device and the 3D printing terminal respectively;

[0033] The image acquisition device is used to acquire images of vascular lesions of the patient and send the images of vascular lesions to the central processing unit;

[0034] The central processing unit is used to perform image segmentation on the vascular lesion image to obtain multiple two-dimensional images, generate three-dimensional lesion data based on the multiple two-dimensional images, generate a vascular intervention surgery plan based on the three-dimensional lesion data and surgical instrument parameters, and send the three-dimensional lesion data to the 3D printing terminal; the multiple two-dimensional images correspond to multiple different tissues or structures; the vascular intervention surgery plan includes the surgical instrument model, usage order and instrument path.

[0035] The 3D printing terminal is used to produce a lesion simulation model based on the three-dimensional lesion data; the surgeon performs in vitro practice based on the vascular intervention surgery plan and the lesion simulation model, and sends the in vitro practice results to the central processor;

[0036] The central processing unit is further configured to adjust the vascular intervention surgery plan according to the in vitro training results.

[0037] Optionally, the lesion simulation model is made of photosensitive resin, silicone, hydrogel and artificial blood vessel materials.

[0038] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects: the present invention performs image segmentation on vascular lesion images before surgery to obtain multiple two-dimensional images; generates three-dimensional data of the lesion based on the multiple two-dimensional images; generates a vascular intervention surgery plan based on the three-dimensional lesion data and surgical instrument parameters; creates a lesion simulation model based on the three-dimensional lesion data; the surgeon conducts in vitro practice based on the vascular intervention surgery plan and the lesion simulation model, and adjusts the vascular intervention surgery plan based on the in vitro practice results. This allows the surgeon to fully understand the patient's lesion condition before surgery, improves the accuracy of the vascular intervention surgery plan, and thus improves the safety of the surgery.

[0039] Figures in the specification

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] FIG1 is a flow chart of a method for preoperative planning of vascular intervention provided by the present invention;

[0042] FIG2 is a schematic diagram of a vascular intervention preoperative planning system provided by the present invention. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] The purpose of the present invention is to provide a preoperative planning method and system for vascular intervention, which can not only intelligently generate surgical plans, but also allow surgeons to practice before the operation, which will greatly reduce the patient's surgical risk, shorten the operation time, and improve the safety of the operation.

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0046] As shown in FIG1 , the method for preoperative planning of vascular intervention provided by the present invention includes:

[0047] Step 100: Obtain images of the patient's vascular lesions. The images of vascular lesions include aneurysms, vascular malformations, vascular stenosis, vascular dissection, and other disease lesions.

[0048] First, the vascular lesion images are preprocessed, including image denoising and contrast enhancement, to improve the image resolution.

[0049] In this embodiment, a Gaussian filter is used to perform denoising on the vascular lesion image. The Gaussian filter performs filtering by performing a convolution operation on a two-dimensional Gaussian function and the image. The formula of the two-dimensional Gaussian function is: G(x, y) = (1 / 2πσ 2 )*e^-(x 2 +y 2 ) / 2σ 2 , where G(x,y) represents the value of the two-dimensional Gaussian function at (x,y), and σ is the standard deviation of the two-dimensional Gaussian function.

[0050] The contrast enhancement processing based on the local area can make the local structure in the vascular lesion image clearer. In this embodiment, the contrast enhancement processing of the vascular lesion image is performed using the following formula: new =I old ×g(x,y);

[0051] Among them, I old For vascular lesion imaging, I new is the enhanced vascular lesion image, (x, y) is the coordinate of the currently processed pixel, and g(x, y) is the coefficient calculated based on the grayscale values ​​of the eight pixels around the pixel (x, y).

[0052] Step 200: Segment the vascular lesion image to obtain multiple two-dimensional images. The multiple two-dimensional images correspond to multiple different tissues or structures. Image segmentation techniques are used to segment the vascular lesion image into different tissue or structural components, including threshold segmentation, region growing, edge detection, and the like.

[0053] Specifically, step 200 includes:

[0054] (21) The vascular lesion image is segmented using Otsu's algorithm to determine the tissue area or structure area in the vascular lesion image and obtain a preliminary image set.

[0055] Otsu's algorithm assumes that the image to be processed consists of two parts: foreground and background, and uses statistical methods to calculate the optimal threshold to maximize the distinction between the foreground and background. Using Otsu's algorithm to determine the optimal threshold t*, the image to be processed can be divided into two regions: foreground and background. The foreground region refers to the region composed of pixels with grayscale values ​​higher than the optimal threshold t* in the image to be processed, which usually represents the target object or structure of interest in the image to be processed. The background region refers to the region composed of pixels with grayscale values ​​lower than the optimal threshold t* in the image to be processed, which usually represents non-target objects or structures of no interest in the image to be processed.

[0056] The division of foreground and background areas is closely related to different tissues or structures. In vascular lesion images, such as X-rays and MRI, the density or signal intensity of different tissues will be different. Using Otsu's algorithm for image segmentation can separate tissues or structures with similar density or signal intensity, thereby helping doctors diagnose and analyze. For example, in X-rays, bones usually have higher grayscale values, while soft tissues usually have lower grayscale values. Threshold segmentation can distinguish between bones and soft tissues.

[0057] Otsu's algorithm first traverses all pixels in the vascular lesion image, counts the number of pixels at each grayscale level, and obtains a grayscale histogram. The optimal threshold t* is determined as follows:

[0058] Using the formula Count the number of pixels at different gray levels; where Ni is the number of pixels at gray level i, and N is the total number of pixels in the vascular lesion image. When the gray level of pixel n is i, N i (n)=1, otherwise N i (n)=0. I is the total number of gray levels.

[0059] Initialize the intra-class variance σ_w2: σ_w2 = 0.

[0060] Initialize the optimal threshold t*: t*=0.

[0061] For each possible threshold t:

[0062] Calculate the within-class variance:

[0063] The number of pixels N0 of category 0 is: The average grayscale μ0 within category 0 is:

[0064] The number of pixels N1 of category 1 is: The average grayscale μ1 within category 1 is:

[0065] The intra-class variance σ_w2_t is calculated using the formula σ_w2_t=w0*w1*(μ0-μ1)2; where w0=N0 / N and w1=N1 / N.

[0066] The inter-class variance σ_b2_t is calculated using the formula σ_b2_t = σ_z2-σ_w2_t; where σ_z2 is the global variance, σ_z2 = ∑ i (i-μ_z)2 / N,

[0067] If σ_b2_t>σ_b2, then update the inter-class variance and optimal threshold: σ_b2=σ_b2_t, t*=t.

[0068] (22) Performing region growing on the preliminary image set to obtain a grown target image set. Specifically, region growing is the process of growing groups of imaging pixels or regions into larger regions. Starting from a set of data seed points, region growing from these points is performed by merging adjacent pixels with similar attributes such as grayscale and anatomical structure to each seed point into this region, especially imaging of some major blood vessels.

[0069] (23) Perform Canny edge detection on the target image set after growth to obtain multiple two-dimensional images. Specifically, dual threshold detection is used to classify pixels into strong edges and weak edges. A high threshold and a low threshold are set. Pixels above the high threshold are considered strong edges, pixels below the low threshold are considered background, and pixels between the two high and low thresholds are considered weak edges. Using the connection operation, weak edge pixels are connected with strong edge pixels to form a complete edge. If a weak edge pixel is adjacent to a strong edge, it is converted into a strong edge.

[0070] Step 300: Generate three-dimensional data of the lesion based on multiple two-dimensional images.

[0071] Specifically, step 300 includes:

[0072] (31) Multiple two-dimensional images are stacked to obtain a preliminary three-dimensional image.

[0073] (32) The preliminary three-dimensional image is subjected to voxelization and point cloud reconstruction in sequence to obtain a preliminary three-dimensional model of the lesion. Voxelization converts the two-dimensional image into a three-dimensional voxel grid, and point cloud reconstruction generates a point cloud model based on the position information of the two-dimensional image.

[0074] (33) Optimizing the preliminary three-dimensional lesion model to obtain final three-dimensional lesion data.

[0075] The process of optimizing the preliminary lesion three-dimensional model includes operations such as eliminating noise, filling holes, and smoothing the surface. Specific methods include voxel filling algorithm, point cloud filtering, mesh simplification algorithm, structured light triangulation, etc.

[0076] Among them, mesh simplification algorithms, such as the Laplacian smoothing algorithm, are used to reduce the complexity of the preliminary lesion three-dimensional model and reduce unnecessary details.

[0077] Voxel filling algorithm: Eliminate gaps in the preliminary lesion 3D model by filling the holes in the voxel grid.

[0078] Structured light triangulation: Calculates the three-dimensional coordinates of each point on the surface of an object based on the projection and reflection principles of structured light.

[0079] Iterative Closest Point (ICP) algorithm: estimates the relative position between two point clouds by iteratively matching point pairs in the point cloud.

[0080] Step 400: Generate a vascular interventional surgery plan based on the three-dimensional lesion data and surgical instrument parameters. The vascular interventional surgery plan includes the surgical instrument models, usage sequence, and instrument path. During the generation of the vascular interventional surgery plan, the order of use of different surgical instrument models can be adjusted. For example, in the order of different coils used for aneurysm packing, after adjusting the model of the first coil, the subsequent coils are automatically adjusted, achieving the purpose of automatically adjusting the surgical plan. The plan can also be automatically adjusted after each step based on the actual surgical situation.

[0081] When the medical device model is known and the hemodynamic data of the medical device is available during the research and development stage, the matching range of the treatment end is formed. The process of forming a vascular interventional surgery plan is the process of determining whether the matching range of the medical device can cover or match the three-dimensional data of the lesion.

[0082] Specifically, step 400 includes:

[0083] (41) The three-dimensional data of the lesion are fused with different types of surgical instruments respectively, and the order of using the different types of surgical instruments is adjusted to obtain a preliminary fused three-dimensional image.

[0084] (42) De-noising and alignment processing are performed on the preliminary fused three-dimensional image in sequence to obtain a corrected fused three-dimensional image.

[0085] In this embodiment, a smoothing filter (e.g., a Gaussian filter) is used to remove noise from the preliminary fused 3D image using the following formula: h(x,y) = ∑[w(k,j)*f(xk,yj)]; where h(x,y) is the pixel value at (x,y) after denoising, f(x,y) is the pixel value at (x,y) in the preliminary fused 3D image, and w(k,j) represents the elements of a weight matrix. Alignment is performed using a coordinate transformation formula, affine transformation, perspective transformation, or rigid body transformation.

[0086] (43) A Canny edge detection algorithm is used to extract edge features of the surgical area in the corrected fused three-dimensional image to obtain a feature-fused three-dimensional image.

[0087] Specifically, the Canny edge detection algorithm uses a formula that includes Gaussian filtering, gradient calculation, non-maximum suppression, and dual threshold detection. Edge features include important anatomical structures and blood vessels in the corrected fused 3D image.

[0088] (44) The ICP algorithm is used to align and match the lesion point cloud and the surgical instrument point cloud in the feature-fused three-dimensional image to obtain an aligned fused three-dimensional image.

[0089] The ICP algorithm involves a least-squares optimization problem between point clouds. The point clouds are aligned by minimizing the mean squared distance between matching point pairs and calculating a transformation matrix. Matching the lesion point cloud and the surgical instrument point cloud involves determining the distance between the lesion and one or more surgical instruments in a 3D display. A narrower distance indicates a higher degree of fit, while a wider distance indicates a higher degree of fit.

[0090] (45) Based on the aligned fused three-dimensional image, geometric calculation and geometric constraints are used to determine the surgical instrument model, usage sequence, and instrument path. For example, the shortest path algorithm can be used to solve the instrument path, and the usage sequence of different types of surgical instruments can be adjusted, such as the sequence of different spring coils used for aneurysm packing. After adjusting the model of the first spring coil, the subsequent spring coils are automatically adjusted to achieve the purpose of automatically adjusting the surgical plan. The plan after each step can be automatically adjusted according to the actual surgical situation.

[0091] Furthermore, the root mean square error (RMSE) was used to evaluate the accuracy of the vascular interventional surgery plan: RMSE = sqrt(∑[d^2] / c); where d is the distance between two point clouds and c is the total number of points in the point cloud.

[0092] Step 500: Produce a lesion simulation model based on the three-dimensional lesion data. In this embodiment, the lesion simulation model is printed or produced using photosensitive resin, silicone, hydrogel, and artificial blood vessel materials to simulate the corresponding characteristics of the lesion.

[0093] The lesion simulation model can be used directly for pre-operative practice, surgical plan formulation and planning, simulate the vascular path and lesions of full-path interventional surgery, and assist surgeons in completing in vitro practice and auxiliary treatment.

[0094] The lesion simulation model includes vascular diseases such as aneurysms, vascular malformations, and vascular stenosis. It has entrances and exits, and special marking points at specific locations for image recognition. It can be connected to an in vitro simulation surgical system to form a complete in vitro vascular circuit, simulating blood and other parameters through the lumen.

[0095] Step 600: Perform in vitro practice or actual surgery based on the vascular intervention surgery plan and the lesion simulation model, and learn and adjust the vascular intervention surgery plan in real time according to the in vitro practice results or the actual surgery situation.

[0096] As shown in Figure 2, the present invention also provides a vascular intervention preoperative planning system, comprising: an image acquisition device 1, a central processing unit 2, and a 3D printing terminal 3. The central processing unit 2 is connected to the image acquisition device 1 and the 3D printing terminal 3 respectively.

[0097] The image acquisition device 1 is used to acquire the patient's vascular lesion image and send the vascular lesion image to the central processor 2.

[0098] The central processing unit 2 is configured to segment the vascular lesion image to obtain multiple two-dimensional images, generate three-dimensional lesion data based on the multiple two-dimensional images, generate a vascular interventional surgery plan based on the three-dimensional lesion data and surgical instrument parameters, and transmit the three-dimensional lesion data to the 3D printing terminal 3. The multiple two-dimensional images correspond to multiple different tissues or structures. The vascular interventional surgery plan includes the surgical instrument model, usage sequence, and instrument path (i.e., surgical steps).

[0099] For example: For intracranial aneurysms, based on their morphology, location and relationship with surrounding blood vessels, vascular interventional surgery plans such as simple embolization, stent-assisted embolization, blood flow guidance device, and balloon-assisted embolization are given, including the order of coil use.

[0100] Furthermore, for simple embolism, if the first coil model is not executed according to the recommended plan, then the system will further give the recommended placement order of the next coil model after collecting the first actual coil model, and the subsequent coil order will also be adjusted in real time until the operation is completed. Similarly, for stent-assisted embolization, if the stent is placed first, then the placement order of the next coils of different models will be given. If the coil is placed first, then the model and position of the stent will be given, and real-time adjustment and real-time guidance will be given for the next treatment plan. During this process, the system will also adjust and self-learn according to the actual surgical plan, so that it can give a better treatment plan the next time a similar artery is encountered.

[0101] In this embodiment, the central processing unit 2 can receive the vascular lesion image sent by the image acquisition device 1, perform 3D reconstruction, obtain a more accurate 3D reconstructed vascular image through manual correction, and then send the reconstructed 3D lesion data to the 3D printing terminal 3. Furthermore, after the 3D reconstruction, the central processing unit 2 generates multiple vascular interventional surgery plans.

[0102] The 3D printing terminal 3 is used to create a lesion simulation model based on the three-dimensional lesion data. The lesion simulation model is made of photosensitive resin, silicone, hydrogel, and artificial blood vessel materials. The surgeon performs in vitro practice or actual surgery based on the vascular interventional surgery plan and the lesion simulation model, and transmits the in vitro practice results or actual surgical results to the central processor 2.

[0103] In this embodiment, the 3D printing terminal 3 is set up in the hospital, and can receive the three-dimensional data of the lesion sent by the central processor 2 for 3D printing, making a lesion simulation model, and printing and forming it in a timely and rapid manner, thereby shortening the patient's waiting time.

[0104] The central processing unit 2 is also used to learn and adjust the vascular interventional surgery plan in real time based on the results of in vitro training or actual surgical conditions. The central processing unit 2 can also receive feedback from in vitro training and actual surgical data to conduct self-learning and optimize the algorithm.

[0105] In addition, the vascular intervention preoperative planning system further includes a display 4. The display 4 is connected to the central processing unit 2 and is used to display the vascular intervention surgery plan and three-dimensional lesion data.

[0106] Specifically, the surgeon conducts in vitro practice on the training terminal based on the vascular interventional surgery plan and the lesion simulation model. The training terminal is equipped with a matching camera 5, which transmits the image to the central processing unit 2, which identifies and matches special markers in the image, generates evaluation data, and guides the surgeon in pre-operative practice. The special markers are physical markers on the lesion simulation model. Through lesion shape feature extraction and image processing, they are matched with pre-defined templates or target features. The surgical image is corrected using geometric transformation methods to align the special markers with the target position.

[0107] In this invention, surgeons can practice or perform real surgeries in vitro, generate realistic solutions, and upload the generated interventional surgery images and text descriptions to the central processing unit 2. The generated interventional surgery plans are then compared with the intelligently generated vascular interventional surgery plans, and improved solutions are provided. The central processing unit 2 then learns and optimizes, enabling it to self-learn the optimization algorithm. This means that reinforcement learning is used to continuously update the vascular interventional surgery plans and lesion simulation models.

[0108] The present invention sends the patient's vascular lesion image to the central processing unit 2 through the sending port of the image acquisition device 1, performs data preprocessing, image segmentation, three-dimensional reconstruction and model optimization, performs manual correction on erroneous or incomplete parts to make the lesion more simulated, and intelligently generates a vascular intervention surgery plan, which is then remotely sent to the hospital's 3D printing terminal 3 via the network to produce a lesion simulation model. The lesion simulation model can be used alone or connected to an in vitro simulation surgery system to form a complete surgical path. Then, based on the intelligently generated vascular intervention surgery plan, in vitro exercises, surgical plan formulation and planning are carried out before the operation, and the plan modifications generated by the in vitro exercises are fed back to the central processing unit 2, so that it can self-learn and optimize the plan algorithm, thereby improving the accuracy of the vascular intervention surgery plan and reducing surgical risks.

[0109] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0110] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for pre-operative planning of vascular intervention, characterized in that, the method for pre-operative planning of vascular intervention includes: obtaining the vascular lesion image of the patient; performing image segmentation on the vascular lesion image to obtain a plurality of two-dimensional images; the plurality of two-dimensional images correspond to a plurality of different tissues or structures; generating lesion three-dimensional data based on the plurality of two-dimensional images; generating a vascular intervention surgical plan based on the lesion three-dimensional data and surgical instrument parameters; the vascular intervention surgical plan includes the surgical instrument model, the usage sequence and the instrument path; manufacturing a lesion simulation model based on the lesion three-dimensional data; performing in vitro practice or performing an actual operation based on the vascular intervention surgical plan and the lesion simulation model, and learning and adjusting the vascular intervention surgical plan in real time according to the in vitro practice results or the actual surgical conditions.

2. The method for pre-operative planning of vascular intervention according to claim 1, characterized in that, before performing image segmentation on the vascular lesion image, the method for pre-operative planning of vascular intervention further includes: performing denoising and contrast enhancement processing on the vascular lesion image in sequence.

3. The method for pre-operative planning of vascular intervention according to claim 2, characterized in that, using a Gaussian filter to perform denoising processing on the vascular lesion image.

4. The method for pre-operative planning of vascular intervention according to claim 1, characterized in that, performing image segmentation on the vascular lesion image to obtain a plurality of two-dimensional images, specifically including: using Otsu's algorithm to segment the vascular lesion image to determine the tissue area or structure area in the vascular lesion image, and obtaining a preliminary image set; performing region growing processing on the preliminary image set to obtain a target image set after growth; performing Canny edge detection on the target image set after growth to obtain a plurality of two-dimensional images.

5. The method for pre-operative planning of vascular intervention according to claim 1, characterized in that, generating lesion three-dimensional data based on a plurality of two-dimensional images, specifically including: stacking a plurality of two-dimensional images to obtain a preliminary three-dimensional image; performing voxelization and point cloud reconstruction processing on the preliminary three-dimensional image in sequence to obtain a preliminary lesion three-dimensional model; optimizing the preliminary lesion three-dimensional model to obtain the final lesion three-dimensional data.

6. The method for pre-operative planning of vascular intervention according to claim 1, characterized in that, generating a vascular intervention surgical plan based on the lesion three-dimensional data and surgical instrument parameters, specifically including: fusing the lesion three-dimensional data with surgical instruments of different models respectively, and adjusting the usage sequence of surgical instruments of different models to obtain a preliminary fused three-dimensional image; performing denoising and alignment processing on the preliminary fused three-dimensional image in sequence to obtain a corrected fused three-dimensional image; using the Canny edge detection algorithm to extract the edge features of the surgical area in the corrected fused three-dimensional image to obtain a feature fused three-dimensional image; using the ICP algorithm to align and match the lesion point cloud and the surgical instrument point cloud in the feature fused three-dimensional image to obtain an aligned fused three-dimensional image; based on the aligned fused three-dimensional image, using geometric calculation and geometric constraints to determine the surgical instrument model, the usage sequence and the instrument path.

7. A pre-operative vascular intervention planning system, characterized in that, the pre-operative vascular intervention planning system comprises: an image acquisition device, a central processor and a 3D printing terminal; the central processor is respectively connected to the image acquisition device and the 3D printing terminal; the image acquisition device is configured to acquire vascular lesion images of a patient and send the vascular lesion images to the central processor; the central processor is configured to perform image segmentation on the vascular lesion images to obtain a plurality of two-dimensional images, generate lesion three-dimensional data based on the plurality of two-dimensional images, generate a vascular intervention surgical plan based on the lesion three-dimensional data and surgical instrument parameters, and send the lesion three-dimensional data to the 3D printing terminal; the plurality of two-dimensional images correspond to a plurality of different tissues or structures; the vascular intervention surgical plan includes surgical instrument models, usage sequences and instrument paths; the 3D printing terminal is configured to fabricate a lesion simulation model based on the lesion three-dimensional data; a surgeon performs in vitro practice or actual surgery based on the vascular intervention surgical plan and the lesion simulation model, and sends the in vitro practice results or actual surgical conditions to the central processor; the central processor is further configured to learn and adjust the vascular intervention surgical plan in real time according to the in vitro practice results or actual surgical conditions.

8. The pre-operative vascular intervention planning system according to claim 7, characterized in that, the lesion simulation model is fabricated using photosensitive resin, silicone, hydrogel and artificial vascular materials.

Images