Method for generating intracranial aneurysm vessel model and computer readable storage medium
By generating intracranial aneurysm vascular models through automated level set segmentation and centerline extraction, the problems of complex manual operation and unstable accuracy in existing technologies are solved, achieving efficient and accurate vascular model generation and expanding the application of hemodynamic simulation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HANGZHOU ARTERYFLOW TECH CO LTD
- Filing Date
- 2022-12-30
- Publication Date
- 2026-07-03
AI Technical Summary
Existing technologies require cumbersome manual operations to generate intracranial aneurysm vascular models, resulting in high time and labor costs. Furthermore, the accuracy of the models is easily affected by subjective factors, limiting the application of hemodynamic simulation.
By acquiring three-dimensional medical images, an intracranial aneurysm vascular model is generated using level set segmentation and centerline extraction. Regions of interest are automatically selected and an initial vascular model is generated, avoiding manual trimming and hole filling operations. The accuracy and efficiency of the model are improved by combining thresholding and level set methods.
It reduces manual processing time and costs, improves model generation speed and accuracy, enhances model robustness, and expands the application scenarios of hemodynamic simulation, including real-time intraoperative applications.
Smart Images
Figure CN115984305B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of medical image processing technology, and in particular to a method for generating an intracranial aneurysm vascular model and a computer-readable storage medium. Background Technology
[0002] With the continuous development of technology, hemodynamic simulation technology has been increasingly widely used in the fields of aneurysm auxiliary diagnosis and surgical planning. Hemodynamic simulation specifically refers to obtaining the dynamic parameters of blood flow and vessel wall by establishing a geometric model of the blood vessel, applying boundary conditions at the vessel inlet and outlet, and then solving the Navier-Stocks equations through numerical calculation methods.
[0003] As is well known, the quality of simulation results depends primarily on the quality of the model. First, the vascular model should be as close as possible to the actual vascular lumen. A model that is too large or too small will introduce simulation errors, rendering the results meaningless. Second, the vascular model must remove small perforating vessels while ensuring reasonable results, thereby controlling the number of meshes and saving computational costs.
[0004] Currently, vascular model reconstruction technology mainly consists of two steps: 1. Obtaining a preliminary vascular model based on vascular medical images (DSA, CTA, or MRA) using image segmentation algorithms (such as thresholding, active contouring, watershed analysis, level set analysis, region growing, etc.) and the traveling cubes algorithm; 2. Performing manual preprocessing operations on the model, such as trimming, filling holes, and smoothing, using third-party software (such as Geomagic). The second step typically consumes significant time and manpower, and the accuracy of the model is easily affected by the subjective factors of the preprocessing operator, thus limiting the clinical application of hemodynamic simulation.
[0005] In general, the process of reconstructing intracranial aneurysm vascular models using traditional methods usually requires complex manual preprocessing of the model, which is time-consuming and labor-intensive, thus restricting the application of hemodynamic simulation technology. Summary of the Invention
[0006] Therefore, it is necessary to provide a method for generating an intracranial aneurysm vascular model to address the aforementioned technical problems.
[0007] The method for generating an intracranial aneurysm vascular model in this application includes:
[0008] Acquire three-dimensional medical images containing intracranial aneurysms;
[0009] The three-dimensional medical image is segmented to obtain a first horizontal set based on the surface of intracranial arteries;
[0010] From the intracranial aneurysm vascular model generated based on the first level set, regions of interest including intracranial aneurysms are selected.
[0011] Extract the centerline and radius along the line from the region of interest. Extracting the centerline includes extracting the centerline of intracranial arteries and extracting the centerline of intracranial aneurysms.
[0012] An initial duct model is generated based on each centerline, and then a second level set based on the surface of the initial duct model is obtained. After segmentation, an intracranial aneurysm vascular model is generated.
[0013] Optionally, the three-dimensional medical image is segmented to obtain a first horizontal set based on the surface of intracranial arteries, specifically including:
[0014] The three-dimensional medical image is segmented and its surface reconstructed sequentially using a thresholding method to obtain an initial level set.
[0015] Using the vessel walls of the three-dimensional medical image as the zero level set, a first level set based on the surface of intracranial arteries is obtained based on the initial level set.
[0016] Optionally, the filtering to obtain the region of interest including the intracranial aneurysm includes: receiving a clipping location specified by the user, performing clipping on the intracranial aneurysm vascular model generated based on the first level set, and obtaining the region of interest including the intracranial aneurysm.
[0017] Optionally, the cutting location includes the proximal and distal openings of the region of interest of the intracranial aneurysm.
[0018] Optionally, extracting the centerline and radius along the line from the region of interest includes: obtaining a Thiessen polygon map of the region of interest, obtaining the centerline of the region of interest based on the Thiessen polygon map, obtaining the inscribed sphere along the line based on the centerline, and obtaining the radius along the line based on the inscribed sphere.
[0019] Optionally, the extraction of the intracranial aneurysm centerline includes: receiving the location of the aneurysm apex specified by the user, and extracting the intracranial aneurysm centerline using the location of the aneurysm apex.
[0020] Optionally, an initial vascular model is generated based on each centerline, and then a second level set based on the surface of the initial vascular model is obtained. After segmentation, an intracranial aneurysm vascular model is generated, specifically including:
[0021] Multiple initial pipe models are generated by sweeping along each centerline;
[0022] Obtain the second level set based on the surface of each initial pipe model;
[0023] The image is segmented using the second level set, and the segmentation results are merged to generate an intracranial aneurysm vascular model.
[0024] Optionally, the method for generating the intracranial aneurysm vascular model further includes: extracting the aneurysm neck plane, obtaining the aneurysm cavity, and then obtaining the hemodynamic parameters of the aneurysm.
[0025] Optionally, the generation method further includes: taking the intersection of data from at least one of the first level set and / or the second level set with the two-dimensional data of each layer of the three-dimensional medical image, and displaying the intersection marker in the three-dimensional medical image.
[0026] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for generating an intracranial aneurysm vascular model as described in this application.
[0027] The method for generating the intracranial aneurysm vascular model in this application has at least the following effects:
[0028] This application avoids manual preprocessing operations such as trimming, patching, and smoothing, saving labor costs. Perforating vessels are automatically deleted when generating the initial pipe model, eliminating the need for manual trimming; the initial pipe model is generated based on each centerline, resulting in a second level set for segmentation, without the need for patching or smoothing.
[0029] The method for generating intracranial aneurysm vascular models provided in this application avoids cumbersome preprocessing operations and improves the generation speed of the model while ensuring the accuracy of the intracranial aneurysm vascular model. Attached Figure Description
[0030] Figure 1 This is a flowchart illustrating a method for generating an intracranial aneurysm vascular model according to an embodiment of this application.
[0031] Figure 2 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0032] In clinical practice, accurate intracranial aneurysm vascular models help doctors make precise diagnoses of patients. Traditional methods for generating intracranial aneurysm vascular models include manual preprocessing. Subjective biases introduced during this process can affect the accuracy of the generated intracranial aneurysm vascular model. Preprocessing also prolongs the time required to generate the intracranial aneurysm vascular model, increasing both time and labor costs.
[0033] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0034] To solve the above technical problems, please refer to Figure 1 One embodiment of this application provides a method for generating an intracranial aneurysm vascular model, including steps S100 to S500. Wherein:
[0035] Step S100: Acquire a three-dimensional medical image containing the intracranial aneurysm. The three-dimensional medical image includes, but is not limited to, DSA, CTA, and MRA images.
[0036] Step S200: Segment the three-dimensional medical image to obtain a first level set based on the surface of intracranial arteries.
[0037] Step S300: From the intracranial aneurysm vascular model generated based on the first level set, select the region of interest that includes the intracranial aneurysm.
[0038] Step S400: Extract the center line and radius along the line for the region of interest. Extracting the center line includes extracting the center line of intracranial arteries and extracting the center line of intracranial aneurysms.
[0039] Step S500: Generate an initial pipeline model based on each centerline, and then obtain a second level set based on the surface of the initial pipeline model. After segmentation, generate an intracranial aneurysm vascular model.
[0040] In each embodiment of this application, the first level set, the second level set, and the initial level set all include a zero level set as the region boundary, which is used to delineate the area inside and outside. During image segmentation, the grayscale difference between the blood vessel contour and other regions in the 3D medical image is used for re-segmentation, so that the zero level set closely overlaps with the vessel wall of the intracranial aneurysm, ultimately obtaining a segmented binary image of the intracranial aneurysm vessel. The model obtained after 3D reconstruction of this image has a smooth surface.
[0041] In this embodiment, a centerline is extracted from the region of interest in the intracranial aneurysm vascular model generated from the first level set. An initial vascular model is then generated based on this centerline, leading to a second level set. This second level set is used for segmentation to generate the intracranial aneurysm vascular model. This also means that this embodiment does not require specific processing of small perforating vessels, as they are automatically deleted when generating the initial vascular model, avoiding the complex process of manually removing perforating vessels.
[0042] The method for generating the intracranial aneurysm vascular model in this embodiment significantly reduces the time and manpower required for reconstruction, and the resulting model is less susceptible to human factors, exhibiting high robustness. This embodiment lowers the technical threshold for hemodynamic simulation, broadens its application scenarios, and makes its application in real-time intraoperative scenarios possible.
[0043] The following section provides a combined description of the optional or alternative sub-steps for different steps.
[0044] Step S200 specifically includes steps S210 to S220.
[0045] Step S210: The three-dimensional medical image is segmented and its surface reconstructed sequentially using the thresholding method to obtain an initial level set.
[0046] Step S220: Using the blood vessel wall of the three-dimensional medical image as the zero level set, a first level set based on the surface of the intracranial artery is obtained based on the initial level set.
[0047] In this embodiment, the first level set is obtained based on the thresholding method, and the second level set is obtained based on the initial pipeline model. Both the level sets obtained by the thresholding method and the level sets obtained by the initial pipeline model are close to the actual intracranial artery contours, and both can save the time consumed by the level set algorithm, thus balancing the accuracy and efficiency of the level set algorithm. Moreover, the time consumed by the thresholding method segmentation is much less than that consumed by the level set method segmentation. The intracranial aneurysm vascular model finally generated in this embodiment improves the generation efficiency while ensuring high accuracy.
[0048] Specifically, the image is first segmented using a thresholding method, and the surface of the segmented image is reconstructed using a moving cube algorithm to obtain an initial level set. The threshold can be a fixed gray value or a fixed percentage, such as 15% of the maximum gray value in the image.
[0049] Because the threshold method is very sensitive to the choice of threshold, the model obtained by the threshold method often differs significantly from the real blood vessel boundary due to the combined effects of objective factors such as contrast agent inhomogeneity and partial volume effect, as well as subjective factors such as the user's cognitive bias of the blood vessel boundary.
[0050] In step S220, the initial level set is constructed based on the results of the first segmentation (segmentation using the threshold method). The level set method is used to iteratively search for the contour with the largest gray value gradient in the image. Then, the image is binarized and reconstructed in three dimensions using this contour as the boundary as the zero level set to obtain the first level set. This results in the reconstructed blood vessel model having higher accuracy and repeatability.
[0051] For step S300, from the first level set, the region of interest including intracranial aneurysms is selected;
[0052] Further, the process of filtering to obtain the region of interest including the intracranial aneurysm includes: receiving a user-specified clipping location, performing clipping on the intracranial aneurysm vascular model generated based on a first level set, and obtaining the region of interest including the intracranial aneurysm. The clipping location includes the proximal and distal openings of the region of interest of the intracranial aneurysm.
[0053] The first level set obtained in the previous section was based on the entire 3D medical image. In this step, the region of interest is obtained through cropping. After receiving the cropping location specified by the user, the proximal end of the vessel of interest and the distal end of its branches are cropped according to the needs of hemodynamic simulation. The vessel of interest includes the segment where the aneurysm lesion is located, as well as adjacent segments and connected major branches, but excludes very small perforating vessels and segments far from the lesion location.
[0054] For step S400, the center line and radius along the line are extracted for the region of interest. Extracting the center line includes extracting the center line of intracranial arteries and extracting the center line of intracranial aneurysms.
[0055] In step S400, extracting the centerline of the intracranial aneurysm includes: receiving the position of the aneurysm top specified by the user, and extracting the centerline of the intracranial aneurysm using the position of the aneurysm top.
[0056] As can be seen from the context, the steps requiring user intervention in each embodiment include: specifying the aneurysm top location and specifying the trimming location. These two simple interactive operations have virtually no impact on the time required to generate the intracranial aneurysm vascular model.
[0057] When selecting the aneurysm dome, for small and medium-sized saccular aneurysms, select one point in the aneurysm dome area for marking; for large and giant aneurysms, select multiple points in the aneurysm dome area for marking; for aneurysms with complex shapes such as horn-shaped or petal-shaped, select points in each aneurysm cavity area for marking; for tandem aneurysms, select points at each aneurysm dome for marking.
[0058] In step S400, the center line and radius along the line are extracted from the region of interest, including: obtaining the Thiessen polygon map of the region of interest, obtaining the center line of the region of interest based on the Thiessen polygon map, obtaining the inscribed sphere along the line based on the center line, and obtaining the radius along the line based on the inscribed sphere.
[0059] Due to physiological parameters, the region of interest typically includes the proximal opening and multiple distal openings. This step calculates the Voronoi diagram (Thyson polygon diagram) from the proximal opening to each distal opening, as well as the Voronoi diagram from the proximal opening to each aneurysm marker.
[0060] Based on the various Voronoi diagrams, we obtain the arrays of the maximum circumscribed sphere radii along the centerline from the proximal opening to the distal openings, and the arrays of the maximum circumscribed sphere radii along the centerline from the proximal opening to the aneurysm markers. This yields the circumscribed sphere along the line and its corresponding radius. The maximum circumscribed sphere is the sphere centered at points along the centerline. The sequence of radii of the maximum circumscribed spheres corresponding to all points along the centerline constitutes the maximum circumscribed sphere radius array.
[0061] Step S500 specifically includes steps S510 to S530. Wherein:
[0062] Step S510: Sweep along each centerline to generate multiple initial pipe models;
[0063] Step S520: Obtain the second level set based on the surface of each initial pipe model;
[0064] Step S530: The image is segmented using the second level set, and the segmentation results are merged to generate an intracranial aneurysm vascular model.
[0065] Specifically, in steps S510 to S530, sweeping is performed along each centerline to generate initial pipe models composed of triangular facets. The diameter variation of the initial pipe model can be uniform or follow a non-uniform distribution. For example, the pipe diameter at each point on the centerline can be equal to the maximum inscribed sphere radius at that point multiplied by a constant (such as 0.9). To improve the efficiency of generating the initial pipe model, the average, minimum, or median of all elements in the maximum inscribed sphere radius array of the entire blood vessel branch (i.e., the inscribed sphere radius along the line) is used as the sampling step size to sample the centerline.
[0066] Steps S520-S530 involve defining initial level sets for the original 3D images of each initial pipe model surface, then segmenting the image using the level set method to obtain a sequence of binary images corresponding to each blood vessel—a binary image sequence. The maximum value of all binary image sequences corresponding to blood vessels is then processed (i.e., the union of the binary image sequences corresponding to each blood vessel is taken) to obtain the final binary image sequence. Finally, the moving cube algorithm is used to reconstruct the surface of this binary image sequence, resulting in a blood vessel model composed of the blood vessels of interest.
[0067] Maximum value processing is performed using the following formula:
[0068] G(x,y)=max(G1(x,y),G2(x,y),...,G n (x, y))
[0069] Where n is the number of pipes, Gm(x,y) is the gray value of one of the pipes at coordinates (x,y) in the binary image sequence, m is 1 to n, and G(x,y) is the gray value of different pipes after merging.
[0070] In one embodiment, the method for generating an intracranial aneurysm vascular model further includes: extracting the aneurysm neck plane, obtaining the aneurysm cavity, and then obtaining the hemodynamic parameters of the aneurysm.
[0071] Specifically, the intracranial aneurysm vessel model and its internal flow domain generated in step S530 are meshed to obtain an unstructured mesh discrete model. After setting boundary conditions at each vessel inlet and outlet, hemodynamic simulation calculations are performed, and parameters such as blood flow velocity, pressure distribution, wall shear stress, oscillating shear factor, and relative residence time are obtained based on the calculation results.
[0072] The aneurysm neck plane is extracted using methods including, but not limited to, manual line-drawing extraction, automated centerline-based extraction, automated vascular skeleton-based extraction, automated extraction based on mother vessel reconstruction, and automated AI-based extraction. After extraction, the aneurysm cavity is obtained based on the extracted aneurysm neck plane, and hemodynamic parameters of the aneurysm cavity are calculated, including blood flow velocity, pressure distribution, wall shear stress, oscillatory shear factor, relative residence time, and rupture similarity fraction.
[0073] In one embodiment, the method for generating an intracranial aneurysm vascular model, in addition to steps S100 to S500, further includes: taking the intersection of data from at least one of the first level set and / or the second level set with the two-dimensional data of each layer of the three-dimensional medical image, and displaying the intersection marker in the three-dimensional medical image.
[0074] This embodiment displays the reconstructed vascular models at each stage, including a first horizontal set based on the surface of intracranial arteries and a second horizontal set based on the surface of each initial duct model, in a semi-transparent spatial overlap with the original 3D image. The coordinates of the intersecting pixels between the reconstructed vascular models and each slice of the 3D image are used to locate the intersecting pixels and highlight them, facilitating observation of the proximity of the highlighted edges of the vascular models to the vascular boundaries in the image. Furthermore, the embodiment includes real-time updates of the intersection markers based on the reconstructed vascular models at different stages. For example, when using thresholding for segmentation, the segmented image and the reconstructed model are updated in real-time by adjusting the threshold.
[0075] It should be understood that, Figure 1Each step in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.
[0076] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 2 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for generating an intracranial aneurysm vascular model. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.
[0077] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:
[0078] Step S100: Obtain a three-dimensional medical image containing an intracranial aneurysm;
[0079] Step S200: Segment the three-dimensional medical image to obtain a first level set based on the surface of intracranial arteries;
[0080] Step S300: From the first level set, filter to obtain regions of interest including intracranial aneurysms;
[0081] Step S400: Extract the center line and radius along the line for the region of interest. Extracting the center line includes extracting the center line of intracranial arteries and extracting the center line of intracranial aneurysms.
[0082] Step S500: Generate an initial pipeline model based on each centerline, and then obtain a second level set based on the surface of the initial pipeline model. After segmentation, generate an intracranial aneurysm vascular model.
[0083] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0084] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered to be within the scope of this specification. When technical features of different embodiments are embodied in the same drawing, it can be regarded as the drawing also disclosing examples of combinations of the various embodiments involved.
[0085] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A method for generating an intracranial aneurysm vascular model, characterized in that, include: Acquire three-dimensional medical images containing intracranial aneurysms; The three-dimensional medical image is segmented to obtain a first horizontal set based on the surface of intracranial arteries; From the intracranial aneurysm vascular model generated based on the first level set, regions of interest including intracranial aneurysms are selected. The centerline and radius along the line are extracted from the region of interest. Extracting the centerline includes extracting the centerline of intracranial arteries and the centerline of intracranial aneurysms. Extracting the centerline of intracranial aneurysms involves receiving the location of the aneurysm's apex as specified by the user and extracting the centerline of the intracranial aneurysm using this location. Specifically, when selecting the aneurysm apex, for large aneurysms, multiple points are selected and marked in the aneurysm apex region; for tandem aneurysms, each aneurysm apex is selected and marked. Multiple initial pipe models are generated by sweeping along each centerline; Obtain the second level set based on the surface of each initial pipe model; The original image is segmented using the second level set, and the segmentation results are merged to generate an intracranial aneurysm vascular model.
2. The method for generating an intracranial aneurysm vascular model according to claim 1, characterized in that, The three-dimensional medical image is segmented to obtain a first level set based on the surface of intracranial arteries, specifically including: The three-dimensional medical image is segmented and surface reconstructed sequentially using a thresholding method to obtain an initial level set. Using the vessel walls of the three-dimensional medical image as the zero level set, a first level set based on the surface of intracranial arteries is obtained based on the initial level set.
3. The method for generating an intracranial aneurysm vascular model according to claim 1, characterized in that, The filtering process for obtaining the region of interest including intracranial aneurysms includes: receiving a user-specified clipping location, performing clipping on the intracranial aneurysm vascular model generated based on the first level set, and obtaining the region of interest including the intracranial aneurysm.
4. The method for generating an intracranial aneurysm vascular model according to claim 3, characterized in that, The cutting locations include the proximal and distal openings of the region of interest of the intracranial aneurysm.
5. The method for generating an intracranial aneurysm vascular model according to claim 1, characterized in that, Extracting the centerline and radius along the line from the region of interest includes: obtaining a Thiessen polygon map of the region of interest; obtaining the centerline of the region of interest based on the Thiessen polygon map; obtaining the inscribed sphere along the line based on the centerline; and obtaining the radius along the line based on the inscribed sphere.
6. The method for generating an intracranial aneurysm vascular model according to claim 1, characterized in that, The method for generating the intracranial aneurysm vascular model further includes: extracting the aneurysm neck plane, obtaining the aneurysm cavity, and then obtaining the hemodynamic parameters of the aneurysm.
7. The method for generating an intracranial aneurysm vascular model according to claim 1, characterized in that, The generation method further includes: taking the intersection of data from at least one of the first level set and / or the second level set with the two-dimensional data of each layer of the three-dimensional medical image, and displaying the intersection marker in the three-dimensional medical image.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method for generating an intracranial aneurysm vascular model according to any one of claims 1 to 7.