Three-dimensional cerebral arteriovenous malformation vessel morphology complexity calculation method and device

Abstract

The invention discloses a three-dimensional cerebral arteriovenous malformation blood vessel morphology complexity calculation method and device, and the method comprises the steps: obtaining image data of cerebral arteriovenous malformation blood vessels in a preset region of a brain, extracting a lesion blood vessel structure in the image data based on a threshold segmentation algorithm, and carrying out the image segmentation and synthesis processing, obtaining a three-dimensional blood vessel model of cerebral arteriovenous malformation blood vessels; based on the three-dimensional blood vessel model, multi-term fractal analysis calculation of cerebral arteriovenous malformation blood vessels is carried out, and multi-scale fractal feature values are obtained; and based on the multi-scale fractal feature value, calculating the morphological complexity of the cerebral arteriovenous malformed blood vessel. A set of systematic quantitative parameter system is constructed through fractal analysis, the morphological characteristics of AVM lesions and the complexity of vascular construction of the AVM lesions can be accurately reflected, the lesion characteristics can be conveniently and rapidly evaluated, and the limitation of relying on subjective judgment in the past is avoided.

Description

Technical Field

[0001] The present invention relates to the field of vascular detection imaging, and particularly to a method and device for calculating the morphological complexity of three-dimensional cerebral arteriovenous malformation blood vessels. Background Art

[0002] The existing methods for evaluating the vascular morphology of cerebral arteriovenous malformations (AVMs) mainly include the following four categories: 1) Traditional imaging evaluation methods: By evaluating the AVM morphology and the characteristics of the surrounding brain tissue, combined with the patient's medical history, the rupture risk is speculated. However, these methods are mainly based on experience and qualitative judgment, lacking quantitative indicators; it is difficult to comprehensively describe the morphological and blood flow characteristics inside the AVM center, lacking a comprehensive description of the structure inside the AVM nidus, and unable to fully reveal the potential characteristics of high-risk lesions. 2) Radiomics analysis: By extracting a large number of features of the image data (such as texture features, morphological features, etc.), a machine learning model is used to predict the AVM rupture risk. However, these methods require a large number of data features, and the data processing is complex; relying on a "black box" model, lacking the ability to explain the physiological and pathological processes, which limits its promotion in clinical practice. 3) Traditional hemodynamic analysis: Techniques such as computational fluid dynamics (CFD) are used to simulate the blood flow characteristics inside the AVM. However, the CFD calculation process is complex and time-consuming, and at the same time depends on an idealized blood flow model, making it difficult to reflect the real dynamic blood flow state. 4) Cerebral angiography parameter analysis: Based on digital subtraction angiography (DSA), blood flow-related parameters are extracted. However, DSA can only provide two-dimensional image information, making it difficult to describe the complex hemodynamic characteristics in three-dimensional space; it is easily interfered by factors such as tissue overlap, resulting in errors. Summary of the Invention

[0003] The purpose of the embodiments of the present invention is to provide a method and device for calculating the morphological complexity of cerebral arteriovenous malformation blood vessels. Through fractal analysis, a systematic quantitative parameter system is constructed, which can accurately reflect the morphological characteristics of the AVM lesion and the complexity of its vascular architecture, has high biological significance and clinical practicability, provides intuitive and quantitative indicators, is convenient for quickly evaluating the lesion characteristics in daily diagnosis and treatment, and avoids the limitations of relying on subjective judgment in the past.

[0004] To solve the above technical problems, the first aspect of the embodiments of the present invention provides a method for calculating the morphological complexity of cerebral arteriovenous malformation blood vessels, including the following steps:

[0005] Obtain the image data of the cerebral arteriovenous malformation blood vessels in the preset area of the brain, extract the lesion vascular structure in the image data based on the threshold segmentation algorithm, and perform image segmentation and synthesis processing to obtain the three-dimensional vascular model of the cerebral arteriovenous malformation blood vessels;

[0006] Based on the three-dimensional vascular model, perform multiple fractal analysis calculations on the blood vessels of the cerebral arteriovenous malformation to obtain multi-scale fractal eigenvalue;

[0007] Based on the multi-scale fractal eigenvalue, calculate the morphological complexity of the blood vessels of the cerebral arteriovenous malformation.

[0008] Further, the performing multiple fractal analysis calculations on the blood vessels of the cerebral arteriovenous malformation to obtain multi-scale fractal eigenvalue includes:

[0009] Perform multiple fractal analysis calculations, multifractal analysis calculations, and porosity calculations on the blood vessels of the cerebral arteriovenous malformation to obtain several fractal dimension values, multifractal analysis generalized dimension values, porosity values, and porosity function b coefficient values.

[0010] Further, the performing multiple fractal analysis calculations on the blood vessels of the cerebral arteriovenous malformation includes:

[0011] Calculate the fractal dimension values of the blood vessels of the cerebral arteriovenous malformation by the box-counting method and the ball-covering method respectively.

[0012] Further, the calculation formula for calculating the fractal dimension value of the blood vessels of the cerebral arteriovenous malformation by the box-counting method is:

[0013]

[0014] where, FD b is the box dimension fractal dimension value, N(ε) is the number of boxes containing the fractal object, and ε is the size of each box.

[0015] Further, the calculation formula for calculating the fractal dimension value of the blood vessels of the cerebral arteriovenous malformation by the ball-covering method is:

[0016]

[0017] where, FD m is the ball dimension fractal dimension value, r is the radius value of each sphere, and N(r) is the number of spheres inside the fractal structure.

[0018] Further, the calculation formula for the porosity value is:

[0019]

[0020] where, ε is the size of the box, σ ε is the standard deviation, μ ε is the average pixel intensity in the box, k is the serial number of the box, and CV ε is the coefficient of variation of the pixels in the box.

[0021] Further, the calculation formula for the b coefficient value of the porosity function is as follows:

[0022]

[0023] where L(ε) is the fitting formula of the porosity function, ε is the size of each box, a, b, and c are hyperbolic fitting parameters, N(ε) is the number of boxes containing the fractal object, Q 1 is the number of pixels in all boxes, Q 2 is the sum of the squares of the number of pixels in all boxes, and p(i,ε) is the number of pixel points in the i-th box.

[0024] Further, the calculation formula for the generalized dimension value of the multifractal analysis is as follows:

[0025]

[0026] where D q is the generalized dimension value, n is the number of boxes, M i is the number of foreground pixels in the i-th box, M 0 is the total number of pixels in the image, q is the weight coefficient, and L is the size of the box.

[0027] Correspondingly, a second aspect of the embodiments of the present invention provides a device for calculating the morphological complexity of cerebral arteriovenous malformation blood vessels, which calculates the morphological complexity of cerebral arteriovenous malformation blood vessels based on the above-mentioned method for calculating the morphological complexity of cerebral arteriovenous malformation blood vessels, including:

[0028] A model construction module, which is used to obtain the image data of cerebral arteriovenous malformation blood vessels in a preset area of the brain, extract the lesion blood vessel structure in the image data based on the threshold segmentation algorithm, and perform image segmentation and synthesis processing to obtain a three-dimensional blood vessel model of the cerebral arteriovenous malformation blood vessels;

[0029] A fractal calculation module, which is used to perform multifractal analysis calculations on the cerebral arteriovenous malformation blood vessels based on the three-dimensional blood vessel model to obtain multi-scale fractal feature values;

[0030] A complexity calculation module, which is used to calculate the morphological complexity of the cerebral arteriovenous malformation blood vessels based on the multi-scale fractal feature values.

[0031] Correspondingly, a third aspect of the embodiments of the present invention provides an electronic device, which is characterized by including: at least one processor; and a memory connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute based on the above-mentioned method for calculating the morphological complexity of cerebral arteriovenous malformation blood vessels.

[0032] Accordingly, a fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the above-mentioned method for calculating the morphological complexity of arteriovenous malformation blood vessels is implemented.

[0033] The above technical solutions of the embodiments of the present invention have the following beneficial technical effects:

[0034] 1. By combining image processing and fractal analysis, the originally irregular morphology of arteriovenous malformation blood vessels that is difficult to accurately measure with the naked eye is transformed into specific quantitative indicators such as fractal dimension values, generalized dimension values, porosity values, etc., realizing a precise description of the blood vessel morphology from macro to micro and from overall to local, providing objective data support for medical research and diagnosis;

[0035] 2. By comprehensively applying multiple fractal analyses, multifractal analyses, and porosity calculations, the characteristics of arteriovenous malformation blood vessels in terms of structural complexity, hierarchical distribution, and spatial relationship with surrounding tissues are explored in all aspects. It can not only locate the key complex parts of the vascular malformation but also understand the density and connectivity of the vascular network, assisting doctors in comprehensively grasping the condition and providing a basis for formulating personalized treatment plans;

[0036] 3. Based on the above precise quantification and revelation of multi-dimensional characteristics, users can anticipate the treatment difficulty and risk in the process of medical decision-making such as surgical planning and interventional treatment strategy selection. For example, judge the surgical resection range and estimate the intraoperative bleeding risk according to the morphological complexity index, so as to optimize the treatment process, improve the treatment success rate, and promote the development of precision medicine in neurosurgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a flowchart of the method for calculating the morphological complexity of arteriovenous malformation blood vessels provided by the embodiments of the present invention;

[0038] Figure 2a is a schematic cross-sectional view of a time-of-flight (TOF) sequence MRI provided by the embodiments of the present invention;

[0039] Figure 2b is a three-dimensional image reconstruction diagram after the vascular system is segmented provided by the embodiments of the present invention;

[0040] Figure 3a is a schematic diagram of the Sierpinski sponge after 1 iteration provided by the embodiments of the present invention;

[0041] Figure 3b is a schematic diagram of the Sierpinski sponge after 2 iterations provided by the embodiments of the present invention;

[0042] Figure 3cIt is a schematic diagram of the 3 - time iteration of the Sierpinski sponge provided by an embodiment of the present invention;

[0043] Figure 4 It is a schematic diagram of the box - counting method (the side length of the cube is 3 pixels and 4 pixels) provided by an embodiment of the present invention;

[0044] Figure 5 It is a schematic diagram of the calculation of the box - counting fractal dimension provided by an embodiment of the present invention;

[0045] Figure 6 It is a schematic diagram of the multifractal spectrum provided by an embodiment of the present invention;

[0046] Figure 7 It is a schematic diagram of the calculation of the b - coefficient of the porosity function provided by an embodiment of the present invention;

[0047] Figure 8a It is a schematic diagram of the rupture of arteriovenous malformation in the superficial part provided by an embodiment of the present invention;

[0048] Figure 8b It is a schematic diagram of the unruptured arteriovenous malformation in the deep part provided by an embodiment of the present invention;

[0049] Figure 9 It is a block diagram of the device for calculating the morphological complexity of the blood vessels of cerebral arteriovenous malformation provided by an embodiment of the present invention.

[0050] Reference numerals:

[0051] 1. Model construction module, 2. Fractal calculation module, 3. Complexity calculation module. Detailed implementation manners

[0052] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the specific implementation manners and with reference to the accompanying drawings. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the descriptions of well - known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present invention.

[0053] Arteriovenous malformation (AVM) is a vascular malformation characterized by abnormal development of arteriovenous in the brain. The blood vessels are intertwined and coiled with each other, and rupture and bleeding may occur, seriously endangering the patient's life. The main treatment methods include: Observation: For AVMs without symptoms or with low risks, regular monitoring can be selected. Surgical resection: Removing the AVM through surgery is one of the methods for radical treatment. Endovascular interventional treatment: Injecting embolizing substances through a catheter to block abnormal blood vessels. Radiosurgery: Using focused radiation (such as gamma knife) to shrink or close the AVM. The annual bleeding risk of untreated AVM is about 1-3%. Through effective treatment, these risks can be greatly reduced. Since AVM is a congenital disease, there is currently no known preventive method. However, through early diagnosis and appropriate treatment, the risk of complications can be reduced and the prognosis can be improved. It is very necessary for accurate identification of high bleeding risk, high surgical risk, and for making auxiliary decisions on endovascular treatment.

[0054] The diagnosis of AVM usually requires imaging examinations, including: Magnetic Resonance Imaging (MRI): It shows the brain structure and blood vessels in detail; Computed Tomography (CT): Especially used when bleeding is suspected; Magnetic Resonance Angiography (MRA) and Computed Tomography Angiography (CTA): Scanning cerebral blood vessels using non-invasive methods, and contrast agents can be injected intravenously to enhance the visualization of cerebral blood vessels in the venous system. Angiography: It shows the blood vessel structure and abnormal connections in detail.

[0055] Please refer to Figure 1 , the first aspect of the embodiment of the present invention provides a method for calculating the morphological complexity of arteriovenous malformation blood vessels in the brain, including the following steps:

[0056] Step S100, obtaining the image data of arteriovenous malformation blood vessels in a preset area of the brain, extracting the lesion blood vessel structure in the image data based on the threshold segmentation algorithm, and performing image segmentation and synthesis processing to obtain a three-dimensional blood vessel model of arteriovenous malformation blood vessels.

[0057] The "preset area" is usually the brain areas where arteriovenous malformation (AVM) is known to have a high incidence, such as the frontal lobe and temporal lobe of the brain, or the suspicious area initially located according to the patient's previous symptoms and other imaging examinations. Images of this area are collected through specific medical imaging devices, such as Magnetic Resonance Angiography (MRA), Computed Tomography Angiography (CTA), etc. These devices use different imaging principles. MRA can clearly show the blood vessel structure based on the effect of the magnetic field on hydrogen protons in the blood; CTA performs X-ray scanning after intravenous injection of contrast agents to highlight the blood vessel images, thereby obtaining the original image data containing arteriovenous malformation blood vessels.

[0058] In an alternative embodiment of the present invention, magnetic resonance cerebral angiography technology is used to scan and image the patient's head using a 3D-TOF sequence. A standard MRI protocol is adopted, including T1-weighted, T2-weighted, and 3D-TOF sequences for cerebral vascular imaging. 3D-TOF imaging is a technique specifically for vascular imaging, based on the Time of Flight principle. In this sequence, an inversion recovery technique is applied to optimize the suppression of static tissue, significantly improving the contrast between vascular signals and background tissue. The specific parameters are: TR = 22 ms, TE = 3.68 ms, flip angle = 18°, and the spatial resolution is 0.52 mm × 0.52 mm. This sequence can clearly display the three-dimensional structure of cerebral blood vessels and is particularly suitable for the precise three-dimensional reconstruction of AVM. The high-resolution image data obtained by 3D-TOF provides a basis for subsequent vascular segmentation and reconstruction, supporting the morphological analysis and diagnosis of the AVM vascular mass.

[0059] Since the characteristics such as the gray value of blood vessels in the image are different from those of the surrounding brain tissue, a suitable gray threshold is set, and the pixel points in the image with gray values higher or lower than this threshold are determined to belong to the vascular structure, so as to initially separate the diseased blood vessels from the complex brain image background. For example, if the blood vessels are shown as brighter areas in the image, a higher brightness threshold is set to extract the vascular pixels corresponding to these bright areas.

[0060] In an alternative embodiment of the present invention, a semi-automatic segmentation technique based on 3D Slicer software, combined with a threshold segmentation algorithm, is used to accurately extract the vascular structure of an AVM (arteriovenous malformation) lesion. First, the original MRI or CT image data is imported into the 3D Slicer software, and the vascular region is extracted using the threshold segmentation algorithm. This algorithm distinguishes the vascular tissue from the background tissue in the image by setting a specific gray threshold. According to the gray values of different vascular tissues, the threshold range can be adjusted manually or automatically to obtain the best segmentation effect and ensure the integrity of the vascular structure. The segmented vascular structure is further 3D reconstructed, and using the 3D visualization function of 3D Slicer, the 2D slice data is synthesized into a complete 3D vascular model. During the reconstruction process, the user can view the reconstruction effect in real time and optimize the result, such as adjusting the viewing angle, magnifying the local area, etc., in order to better present the spatial distribution of the blood vessels. Next, the 3D vascular model is trimmed using the trimming tool in 3DSlicer to remove irrelevant or non-lesion areas. The trimming operation can be performed by setting a trimming plane or manually selecting the region of interest. Finally, the processed 3D vascular structure is standardized to a cross-section of 512×512 pixels, and the size in the vertical direction is adjusted according to the actual size of the AVM lesion to ensure that the resolution and details of the image meet the analysis requirements. Through the above method, the vascular structure of the AVM lesion can be accurately extracted and reconstructed, providing a reliable 3D model for further clinical analysis and surgical planning. Figure 2a shows the original image of an example patient, Figure 2b shows the effect after 3D reconstruction and segmentation of the vascular malformation lesion of an example patient, where the red part is the feeding artery of the patient's malformation mass, the blue part is the draining vein of the patient's malformation mass, and the purple part is the patient's arteriovenous malformation mass.

[0061] In addition, the image segmentation refinement operation also includes removing some misjudged isolated noise points, filling in possible small cavities inside the blood vessels, etc., to improve the accuracy of vascular structure extraction.

[0062] The segmented image data is binarized to convert it into a three-dimensional array corresponding to the image data. Specifically, the image data after blood vessel segmentation is imported into a Python program in NIfTI format. As a standard format in medical image processing, NIfTI format can effectively retain the spatial information and data details of the segmented image. The imported image data is loaded through an image processing library (nibabel) in Python. After loading, the original image is usually presented in the form of a grayscale image, where different grayscale values represent different tissue types. For the convenience of subsequent processing, the image needs to be binarized. Binarization is achieved by setting a threshold, which divides the pixel values in the image into two categories: the part above the threshold is assigned a value of 1 (representing the blood vessel area), and the part below the threshold is assigned a value of 0 (representing the non-blood vessel area). The specific binarization process is realized by judging each pixel one by one using the numpy library in Python. The binarized image data is converted into a three-dimensional array, where each element represents a pixel value in the image, 1 represents blood vessel tissue, and 0 represents the background or non-blood vessel tissue. The size of this three-dimensional array is the same as that of the original image, but its data only contains binary information, which is convenient for subsequent image analysis and processing. Finally, the processed binary image can be used as input data for model training or blood vessel analysis, and is further applied to three-dimensional reconstruction, blood vessel segmentation evaluation, or detailed analysis of AVM lesions.

[0063] Step S200: Based on the three-dimensional blood vessel model, perform multi-fractal analysis and calculation of the cerebral arteriovenous malformation blood vessels to obtain multi-scale fractal feature values.

[0064] Fractal theory is applicable to describing objects with self-similarity and irregular complex structures. The branching and tortuous morphology of cerebral arteriovenous malformation blood vessels exhibit similar complexity at different scales. Based on the three-dimensional blood vessel model and the three-dimensional array, fractal dimension calculations are performed, such as the box dimension method and the sphere dimension method. Taking the box dimension method as an example, "boxes" with different side lengths are used to cover the blood vessel structure, and the number of boxes required for coverage is counted. As the side length of the box changes, there will be a specific power-law relationship between the number of boxes and the side length, and the exponent obtained by fitting the curve is the fractal dimension. Different fractal analysis methods reveal the complexity of blood vessels from different perspectives. For example, the information dimension takes into account the amount of information in the pixel point distribution and is more sensitive to the heterogeneity of the blood vessel structure.

[0065] Due to the different morphological manifestations of cerebral arteriovenous malformation vessels at different scales (from tiny capillary branches to thick feeding arteries and draining veins), a series of fractal characteristic values are obtained through fractal analysis at multiple scales. These values reflect the complexity of the vessels at different levels of detail. For example, the fractal characteristic values at small scales reflect the fine tortuosity of the capillary network, and the characteristic values at large scales reflect the complexity of the branching morphology and orientation of the main vessels. Collectively, they provide rich information for a comprehensive assessment of the vessel morphology.

[0066] Step S300: Calculate the morphological complexity of cerebral arteriovenous malformation vessels based on the multi-scale fractal characteristic values.

[0067] Based on the multi-scale fractal characteristic values, the morphological complexity of cerebral arteriovenous malformation vessels is quantified through specific mathematical models or algorithms. This may involve methods such as weighted summation of fractal characteristic values at different scales and construction of complexity functions. For example, higher weights are given to the small-scale characteristic values because the complexity of the capillary network has a greater impact on physiological functions such as overall hemodynamics. Then, a comprehensive value is calculated based on the weight distribution as an index of morphological complexity. This index can be used to compare the degrees of vascular malformations among different patients, evaluate the severity of the condition, and also provide a key basis for formulating subsequent treatment plans, such as surgical planning and selection of interventional treatment strategies, to help users predict the treatment difficulty and risks.

[0068] The present invention utilizes image processing technology and fractal theory to transform the morphological complexity of cerebral arteriovenous malformation vessels, which is difficult to accurately quantify by the naked eye, into computable and comparable numerical indicators, playing an important role in promoting precision medicine in neurosurgery.

[0069] Specifically, the multiple fractal analysis calculations of cerebral arteriovenous malformation vessels in step S200 to obtain multi-scale fractal characteristic values include:

[0070] Step S210: Conduct multiple fractal analysis calculations, multifractal analysis calculations, and porosity calculations on cerebral arteriovenous malformation vessels to obtain several fractal dimension values, multifractal analysis generalized dimension values, porosity values, and porosity function b coefficient values.

[0071] Among them, multiple fractal analysis is based on fractal geometry theory to explore the various fractal characteristics presented by the object in different local regions. For cerebral arteriovenous malformation vessels, they are not uniform and regular structures, and the branching morphology, density, and tortuosity of blood vessels vary in different parts. Through multiple fractal analysis, these local differences can be deeply explored. For example, in the area close to the malformation core, the blood vessels may show a highly dense and complex branching state, with a relatively high fractal dimension; while in the marginal transition area, the blood vessel branches gradually become sparse, and the fractal dimension decreases accordingly.

[0072] Compared with conventional fractal analysis, multifractal analysis is more sophisticated, taking into account that the distribution of fractal characteristics of objects at different scales is not uniform. Cerebral arteriovenous malformations have a multi-scale hierarchical structure from tiny capillaries to large blood supply and drainage vessels. Multifractal analysis can capture the complexity changes brought about by this hierarchical difference. By constructing a multifractal spectrum and calculating the generalized dimension value, the generalized dimension can quantify the distribution law of singular points of different intensities (corresponding to local points of different complexity in the vascular structure, such as complex branch points, sharp turning points of blood vessels, etc.). For example, a low generalized dimension value area may represent a relatively regular and simple vascular segment, while a high generalized dimension value area highlights a highly irregular, complex and changeable vascular structure aggregation area, which provides a powerful clue for accurately locating the most complex and critical parts of vascular malformations.

[0073] Porosity is analogous to the "gaps" between vascular networks in the study of brain arteriovenous malformations. From a macroscopic perspective, there are gaps similar to pores between the malformed vascular mass and the surrounding normal brain tissue; at a microscopic level, there are also gaps between vascular branches and between vascular walls and internal blood flow. Calculating the porosity value can, on the one hand, help us understand the tightness of the vascular structure. Low porosity means that the blood vessels are densely interwoven, and vice versa. On the other hand, the b coefficient value in the porosity function has special significance, which is related to the connectivity and uniformity of the pores. A higher b coefficient value may indicate that the pores are unevenly distributed, with local concentration or poor connectivity, which is crucial to understanding the perfusion, shunting, and potential risk of thrombosis of blood flow in malformed blood vessels, because the uneven pore structure will affect hemodynamics and trigger a series of pathophysiological changes.

[0074] The fractal dimension values, generalized dimension values, porosity values ​​and porosity function b coefficient values ​​obtained by combining the above calculations constitute a multi-dimensional feature data set, which fully reflects the complex characteristics of cerebral arteriovenous malformations from micro to macro, from local to overall morphology, structure and relationship with surrounding tissues.

[0075] Furthermore, the multiple fractal analysis calculations of the cerebral arteriovenous malformation in step S210 include:

[0076] Step S211, calculating the fractal dimension value of the cerebral arteriovenous malformation by using the box count method and the sphere covering method respectively.

[0077] like Figure 4As shown, the three-dimensional space containing the brain arteriovenous malformation vessels is divided into small cubes (boxes) of equal size by the box-counting method. The side length of these boxes can be gradually reduced, starting from a relatively large initial scale and continuously subdividing the space. For each scale, the number of boxes required to cover the vascular structure is counted. As the side length of the boxes decreases, the number of boxes required to cover the vessels increases according to a specific pattern. Because the vessels have complex branches and tortuous shapes, when the boxes become smaller, more boxes are needed to accurately cover the vessels. For example, at a relatively large scale, when covering with boxes with a side length of 1 cm, perhaps only 100 boxes are needed to roughly cover the main vascular parts; but when the side length of the boxes is reduced to 0.1 cm, maybe 10,000 boxes are required because at this time, more fine vascular branches and tortuous parts need to be captured. The calculation process is as follows: Let the side length of the box be ε, and the number of boxes required to cover the vessels be N(ε). According to fractal theory, there is a power-law relationship N(ε) ∝ ε -d , where d is the fractal dimension. By fitting the data of multiple groups of different ε values and their corresponding N(ε) values, such as using linear regression techniques like the least squares method, and finding the absolute value of the slope, the fractal dimension value of the brain arteriovenous malformation vessels can be obtained. This fractal dimension value reflects the comprehensive characteristics of the vessels' filling, extension, and branching complexity in space. It is not a geometric dimension in the traditional sense but an index quantifying irregularity. Generally speaking, the higher the fractal dimension, the more complex the vascular structure, the more numerous the branches, and the higher the degree of tortuosity.

[0078] By the ball-covering method, the covering unit changes from a cube to a sphere. In three-dimensional space, spheres of different radii are used to cover the brain arteriovenous malformation vessels. Due to the different geometric characteristics of the sphere from the cube, it presents a different scenario when covering irregular objects. When the sphere rolls close to the vessel surface for covering, its fitting degree to the curved and concave parts of the vessel is different from that of the cube, and it can capture the morphological details of the vessel from different angles. Similarly, starting from a relatively large initial radius, gradually reduce the radius of the sphere, and record the number of spheres required to cover the vessels at each radius. The calculation process: Let the radius of the sphere be r, and the number of spheres required to cover the vessels be M(r). Similar to the box-counting method, the two follow a power-law relationship M(r) ∝ r -d' , where d' is the fractal dimension obtained by the ball-covering method. By fitting and analyzing a series of r and M(r) values, d' is calculated. The fractal dimension value obtained by the ball-covering method complements the box-counting method because the covering characteristics of the sphere may highlight some vascular morphological details ignored by the box-counting method. For example, at sharp turns of the vessels and the tips of fine branches, the sphere can fit more naturally, thus obtaining more accurate fractal dimension information reflecting the local complexity, making the assessment of the overall complexity of the brain arteriovenous malformation vessels more comprehensive and three-dimensional.

[0079] By calculating the fractal dimension values through these two methods respectively, not only can the accuracy of the results be verified with each other, but more importantly, it is possible to comprehensively understand the irregular and complex morphological characteristics of the arteriovenous malformation blood vessels from different geometric coverage perspectives, providing a solid data basis for subsequent in-depth analysis of multiple fractal characteristics of the blood vessels and ultimately accurately measuring their morphological complexity.

[0080] Specifically, the calculation formula for the fractal dimension value of arteriovenous malformation blood vessels by the box-counting method is:

[0081]

[0082] Among them, FD b is the box dimension fractal dimension value, N(ε) is the number of boxes containing the fractal object, and ε is the size of each box.

[0083] Specifically, the calculation formula for the fractal dimension value of arteriovenous malformation blood vessels by the ball-covering method is:

[0084]

[0085] Among them, FD m is the sphere dimension fractal dimension value, r is the radius value of each sphere, and M(r) is the number of spheres inside the fractal structure.

[0086] Specifically, the calculation formula for the porosity value in step S300 is:

[0087]

[0088] Among them, ε is the size of the box, σ ε is the standard deviation, μ ε is the average pixel intensity inside the box, k is the serial number of the box, and CV ε is the coefficient of variation of the pixels inside the box.

[0089] Furthermore, porosity is a parameter used to quantify the shape space filling characteristics. Shapes that fill the space more completely have lower porosity, while shapes that fill less space have higher porosity. The porosity value can reflect the distribution of internal and external voids in an object, usually achieved by calculating the change in mass distribution at different scales. On this basis, the calculation method of porosity can be further associated with fractal parameters such as lacunarity to describe the void characteristics of the shape. Lacunarity evaluates the complexity of the shape by quantitatively analyzing the size and distribution of voids. The calculation of porosity involves calculating the porosity rate through box counting at different scales and offsets. This process calculates the coefficient of variation of mass for each box and then obtains the parameter b value of the lacunarity function. Specifically, porosity is represented by calculating the average value of the mass coefficients of a series of boxes, while lacunarity reflects the porosity distribution characteristics of the shape.

[0090] The calculation formula for the b coefficient value of the porosity function is:

[0091]

[0092] where L(ε) is the fitting formula of the porosity function, ε is the size of each box, a, b, c are hyperbola fitting parameters, N(ε) is the number of boxes containing the fractal object, Q 1 is the number of pixels in all boxes, Q 2 is the sum of the squares of the number of pixels in all boxes, and p(i,ε) is the number of pixel points in the i-th box.

[0093] b is the fitting parameter, representing the concavity of the hyperbola. A lower b value corresponds to a wider concavity (i.e., higher lacunarity), indicating that the pore distribution of the object is more dispersed and has larger voids. The b value calculated by the curve fitting method is the lacunarity parameter (blac), which can be used to quantify the pore distribution characteristics of the object.

[0094] Furthermore, multifractal analysis (MFA) describes the characteristics of geometric multifractals through its generalized dimension D q Different from single fractal analysis, multifractal analysis can describe local fractal characteristics through different weights, thus providing more detailed morphological information. Multifractal analysis is widely used in the analysis of three-dimensional medical signals to reveal the fractal characteristics in morphologically complex structures. The generalized dimension D q is the core parameter of multifractal analysis, and its calculation is based on the method of covering a binary cube with boxes of size L. The calculation formula for the generalized dimension value of multifractal analysis is:

[0095]

[0096] Among them, D q is the generalized dimension value, n is the number of boxes, M i is the number of foreground pixels in the i-th box, M 0 is the total number of pixels in the image, q is the weight coefficient, and L is the size of the box.

[0097] M 0 is the total number of pixels in the entire image at a certain scale ∈, which can be used to normalize the number of pixels in each box so that the calculation result is not affected by the image size. M i represents the number of foreground pixels in the i-th box (for example, the number of pixels with a value of 1), and is used to analyze the fractal characteristics of the image.

[0098] M i is the number of pixels in the i-th box, M = ∑M i is the total number of pixels, q is the weight exponent, and its value range is usually within a specified interval. The generalized dimension spectrum consists of a set of fractal dimensions, including: D 0 (when q = 0), which is called the capacity dimension and reflects the space-filling characteristics of the geometric object; D 1 (when q = 1), which is called the information dimension and characterizes the uniformity of information distribution; D 2 (when q = 2), which is called the correlation dimension and represents the correlation between pixels. For a multifractal structure, D q varies with the change of q, while for a monofractal structure, D q is independent of q, that is, it satisfies the condition of D q = D.

[0099] The present invention quantitatively describes the relationship between the structural complexity and the morphological characteristics through the calculation of fractal analysis parameters. The following takes the Sierpinski sponge as an example (such as Figure 3a , Figure 3b and Figure 3c ), and respectively shows the fractal dimension calculation model and related parameters. The Sierpinski sponge is generated at a resolution of 27×27×27 pixels, and its generation rule is: taking Figure 3A as the initial model, Figure 3A is a cube, and the center and the central regions of each face are removed. In each iteration, the size of A is reduced proportionally to 1 / 3 of the original size, and this process is repeated 20 times. Theoretically, the Hausdorff fractal dimension (FD) can be calculated by the formula log 3 20, and the result is 2.727. In Figure 3aAmong them, the fractal analysis parameters are: FDb = 2.656, FDm = 2.714, total number of pixels = 19683, number of non-empty pixels = 14580, number of empty pixels = 5103, ratio of empty pixels to total pixels = 0.259, porosity value (lacunarity) = 0.277, blac = 1577.585. In Figure 3b Among them, the fractal analysis parameters are: FDb = 2.668, FDm = 2.718, total number of pixels = 19683, number of non-empty pixels = 10800, number of empty pixels = 8883, ratio of empty pixels to total pixels = 0.451, porosity value = 0.633, blac = 9.347. In Figure 3c Among them, the fractal analysis parameters are: FDb = 2.682, FDm = 2.738, total number of pixels = 19683, number of non-empty pixels = 8000, number of empty pixels = 11683, ratio of empty pixels to total pixels = 0.594, porosity value = 0.721, blac = 1.682. By comparing and analyzing Figure 3a , Figure 3b and Figure 3c , it is found that both FDb and FDm are close to the theoretical Hausdorff fractal dimension. The ratio of empty pixels to total pixels and the porosity value show a consistent changing trend. As the number of iterations increases, the structure becomes more refined and complex, the porosity gradually increases, and the b coefficient value gradually decreases, fully demonstrating the variation law of the morphological complexity during the continuous evolution of the fractal structure.

[0100] In the specific implementation process of the present invention, a series of medical image data of patients with brain arteriovenous malformations are used as examples. The image data is three-dimensionally reconstructed and segmented, and then input into the analysis program designed by the present invention, and a variety of fractal parameters are calculated, including box dimension (as shown in Figure 5 ), spherical dimension, multifractal analysis (as shown in Figure 6 ), porosity value, b coefficient in the porosity function (as shown in Figure 7 ). By analyzing the multi-scale geometric features of the image, the program successfully extracts the fractal features of different local regions, revealing the complex structures and heterogeneities existing in the image. The calculation results show that the proposed method can not only effectively capture the complex morphological features in medical images, but also reveal the multi-level geometric structures of the images at different scales. By analyzing the images of real patients, we verify the effectiveness of the algorithm in medical image analysis, demonstrating its feasibility and superiority in practical applications. This method provides a new tool for the quantitative analysis of medical images, has strong clinical application potential, and can provide important quantitative support and decision-making basis for the early diagnosis of diseases, lesion evaluation, and treatment plan optimization. Figure 8a and Figure 8bThe imaging results and calculation results of ruptured and unruptured patients are respectively shown. To more intuitively show the morphological differences between the two groups, Figure 8a shows a ruptured arteriovenous malformation with a plexus located superficially and without deep venous drainage or aneurysm. Figure 8b Then it shows an unruptured arteriovenous malformation located in a deeper position and accompanied by deep venous drainage. The sizes of the two plexuses are similar. The calculation results of morphological parameters show that Figure 8a the arteriovenous malformation in Figure 8b has a higher fractal dimension (FD), arteriovenous ratio (A / V), and a lower b coefficient value. Based on the three-dimensional reconstructed image of the plexus,

[0101] Correspondingly, the second aspect of the embodiments of the present invention provides a device for calculating the morphological complexity of arteriovenous malformation blood vessels in the brain, which calculates the morphological complexity of arteriovenous malformation blood vessels in the brain based on the above-mentioned method for calculating the morphological complexity of arteriovenous malformation blood vessels in the brain, including:

[0102] A model construction module 1, which is used to obtain the image data of arteriovenous malformation blood vessels in a preset area of the brain, extract the lesion blood vessel structure in the image data based on the threshold segmentation algorithm, and perform image segmentation and synthesis processing to obtain a three-dimensional blood vessel model of arteriovenous malformation blood vessels;

[0103] A fractal calculation module 2, which is used to perform multi-fractal analysis calculations on arteriovenous malformation blood vessels based on the three-dimensional blood vessel model and three-dimensional array to obtain multi-scale fractal eigenvalue;

[0104] A complexity calculation module 3, which is used to calculate the morphological complexity of arteriovenous malformation blood vessels based on the multi-scale fractal eigenvalue.

[0105] Correspondingly, the third aspect of the embodiments of the present invention provides an electronic device, which is characterized by including: at least one processor; and a memory connected to at least one processor; wherein, the memory stores instructions executable by at least one processor, and the instructions are executed by at least one processor to enable at least one processor to execute based on the above-mentioned method for calculating the morphological complexity of arteriovenous malformation blood vessels in the brain.

[0106] Correspondingly, the fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the above-mentioned method for calculating the morphological complexity of arteriovenous malformation blood vessels in the brain is implemented.

[0107] An embodiment of the present invention aims to protect a method for calculating the morphological complexity of three-dimensional cerebral arteriovenous malformation blood vessels, which includes the following steps: obtaining image data of cerebral arteriovenous malformation blood vessels in a preset area of the brain, extracting the lesion blood vessel structure in the image data based on a threshold segmentation algorithm, and performing image segmentation and synthesis processing to obtain a three-dimensional blood vessel model of cerebral arteriovenous malformation blood vessels; based on the three-dimensional blood vessel model, performing multiple fractal analysis calculations on cerebral arteriovenous malformation blood vessels to obtain multi-scale fractal feature values; based on the multi-scale fractal feature values, calculating the morphological complexity of cerebral arteriovenous malformation blood vessels. The above technical solution has the following effects:

[0108] 1. By combining image processing and fractal analysis, the originally irregular and difficult-to-precisely-measure morphology of cerebral arteriovenous malformation blood vessels is transformed into specific quantitative indicators such as fractal dimension values, generalized dimension values, porosity values, etc., realizing a precise description of the blood vessel morphology from macro to micro and from overall to local, providing objective data support for medical research and diagnosis;

[0109] 2. By comprehensively applying multiple fractal analysis, multifractal analysis, and porosity calculation, the characteristics of cerebral arteriovenous malformation blood vessels in terms of structural complexity, hierarchical distribution, and spatial relationship with surrounding tissues are explored in all aspects. It can not only locate the key complex parts of vascular malformations but also understand the density and connectivity of the blood vessel network, assisting doctors in comprehensively grasping the condition and providing a basis for formulating personalized treatment plans;

[0110] 3. Based on the above precise quantification and revelation of multi-dimensional characteristics, users can anticipate the treatment difficulty and risk in the process of medical decision-making such as surgical planning and interventional treatment strategy selection. For example, judging the surgical resection range and estimating the intraoperative bleeding risk based on the morphological complexity index, thereby optimizing the treatment process, improving the treatment success rate, and promoting the development of precision medicine in neurosurgery.

[0111] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0112] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one or more flows and / or one or more blocks in the flow. Figure 1 in one or more flows and / or one or more blocks Figure 1 of the functions specified in the block or blocks.

[0113] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one or more flows and / or one or more blocks in the flow. Figure 1 in one or more flows and / or one or more blocks Figure 1 of the functions specified in the block or blocks.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows and / or one or more blocks in the flow. Figure 1 in one or more flows and / or one or more blocks Figure 1 of the functions specified in the block or blocks.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the specific embodiments of the present invention, and any modification or equivalent substitution that does not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A method for calculating the morphological complexity of three-dimensional cerebral arteriovenous malformations, characterized in that: The steps include: Acquire image data of cerebral arteriovenous malformation blood vessels in a preset brain area, extract the lesion vascular structure in the image data based on a threshold segmentation algorithm, and perform image segmentation and synthesis processing to obtain a three-dimensional vascular model of the cerebral arteriovenous malformation blood vessels; Based on the three-dimensional blood vessel model, multiple fractal analysis calculations are performed on the cerebral arteriovenous malformation blood vessels to obtain multi-scale fractal eigenvalues; Based on the multi-scale fractal eigenvalues, the morphological complexity of the cerebral arteriovenous malformation is calculated.

2. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 1, characterized in that: The performing of multiple fractal analysis calculations on the cerebral arteriovenous malformation to obtain multi-scale fractal eigenvalues ​​includes: Multiple fractal analysis calculations, multifractal analysis calculations, and porosity calculations are performed on the cerebral arteriovenous malformation blood vessels to obtain a number of fractal dimension values, multifractal analysis generalized dimension values, porosity values, and porosity function b coefficient values.

3. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 2, characterized in that: The performing of multiple fractal analysis calculations of the cerebral arteriovenous malformation blood vessels includes: The fractal dimension values ​​of the cerebral arteriovenous malformation are calculated respectively by the box number method and the sphere covering method.

4. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 3, characterized in that: The calculation formula for calculating the fractal dimension value of the cerebral arteriovenous malformation by the box number method is: Among them, FD b is the box dimension fractal dimension value, N(ε) is the number of boxes containing the fractal object, and ε is the size of each box.

5. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 3, characterized in that: The calculation formula for calculating the fractal dimension value of the cerebral arteriovenous malformation by the sphere covering method is: Among them, FD m is the fractal dimension value of the sphere dimension, r is the radius of each sphere, and N(r) is the number of spheres inside the fractal structure.

6. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 2, characterized in that: Lacunarity ε The calculation formula is: Where ε is the size of the box, σ ε is the standard deviation, μ ε is the average pixel intensity in the box, k is the serial number of the box, CV ε is the coefficient of variation of the pixels within the box.

7. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 2, characterized in that: The calculation formula of the porosity function b coefficient value is: Among them, L(ε) is the porosity function fitting formula, ε is the size of each box, a, b, c are hyperbola fitting parameters, N(ε) is the number of boxes containing fractal objects, Q1 is the number of pixels in all boxes, Q2 is the sum of the squares of the number of pixels in all boxes, and p(i,ε) is the number of pixels in the i-th box.

8. The method for calculating the morphological complexity of brain arteriovenous malformations according to claim 2, characterized in that: The calculation formula of the generalized dimension value of the multifractal analysis is: Among them, D q is the generalized dimension value, n is the number of boxes, M i is the number of foreground pixels in the i-th box, M0 is the total number of pixels in the image, q is the weight coefficient, and L is the size of the box.

9. A device for calculating the morphological complexity of brain arteriovenous malformations, characterized in that: Calculating the morphological complexity of a brain arteriovenous malformation based on the method for calculating the morphological complexity of a brain arteriovenous malformation according to any one of claims 1 to 8 comprises: A model building module, which is used to obtain image data of cerebral arteriovenous malformations in a preset area of ​​the brain, extract the lesion vascular structure in the image data based on a threshold segmentation algorithm, and perform image segmentation and synthesis processing to obtain a three-dimensional vascular model of the cerebral arteriovenous malformations; A fractal calculation module, which is used to perform multiple fractal analysis calculations of the cerebral arteriovenous malformation blood vessels based on the three-dimensional blood vessel model and the three-dimensional array to obtain multi-scale fractal eigenvalues; A complexity calculation module is used to calculate the morphological complexity of the cerebral arteriovenous malformation based on the multi-scale fractal eigenvalue.

10. An electronic device, characterized in that: include: at least one processor; And a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the method for calculating the morphological complexity of brain arteriovenous malformations as described in any one of claims 1-8.

Images