Aneurysm rupture risk prediction system based on medical imaging and hemodynamics

Abstract

This invention relates to the fields of medical imaging and hemodynamics, and discloses an aneurysm rupture risk prediction system based on medical imaging and hemodynamics. The system includes a comprehensive management module, a morphological assessment module, a fluid dynamics assessment module, a spatiotemporal alignment module, and a weighted integration module. This technology uses dual-channel processing to extract the morphological features and hemodynamic parameters of the aneurysm, respectively. The spatiotemporal alignment module achieves precise matching of the two types of information within the same model and time phase, while the weighted integration module fuses and calculates the aligned multi-dimensional risk scores through a preset mechanism. This system overcomes the shortcomings of traditional methods, such as data isolation and reliance on subjective synthesis, enabling correlation analysis between structural abnormalities and pathogenic blood flow patterns, thereby automatically generating a more accurate and objective comprehensive rupture risk score.

Description

Technical Field

[0001] This invention relates to the fields of medical imaging and hemodynamics, specifically to an aneurysm rupture risk prediction system based on medical imaging and hemodynamics. Background Technology

[0002] Currently, clinical assessment of the risk of intracranial aneurysm rupture primarily relies on morphological indicators based on medical imaging. These indicators are synthesized using clinical scales to provide a reference for treatment decisions. Simultaneously, computational fluid dynamics (CFD) techniques can simulate blood flow within the aneurysm, providing dynamic parameters such as wall shear force and oscillatory shear index, which can be used to analyze the mechanical effects of blood flow on the vessel wall. These two methods reveal the potential risks of aneurysms from different perspectives.

[0003] Existing technical solutions typically treat morphological assessment and hemodynamic assessment as two separate parallel processes. Risk assessment either relies on subjective synthesis of the two types of results based on clinical experience or only employs simple parametric comparison or statistical correlation analysis. This approach has drawbacks: the static geometric features of morphology are separated from the dynamic spatiotemporal parameters of hemodynamics, making it impossible to accurately establish a real-time correspondence between specific local morphological abnormalities and their corresponding specific blood flow impact patterns. Isolated data makes it difficult for assessments to reveal the core pathological mechanisms of "how abnormal blood flow leads to morphological changes" and "how specific morphology guides blood flow deterioration," affecting the accuracy of risk prediction and the depth of mechanistic explanation.

[0004] A technical solution is needed to address the disconnect between morphological and hydrodynamic data. The key lies in achieving accurate fusion and mapping of these two heterogeneous data sets within a unified 3D model and cardiac cycle, and on this basis, developing a decision-making mechanism that can objectively and quantitatively integrate multi-dimensional risk information to generate more reliable comprehensive risk assessment results. Summary of the Invention

[0005] The purpose of this invention is to provide an aneurysm rupture risk prediction system based on medical imaging and hemodynamics, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides an aneurysm rupture risk prediction system based on medical imaging and hemodynamics, the system comprising:

[0007] The integrated risk assessment framework management module is used to build and coordinate the morphological data processing channel and the fluid dynamics data processing channel;

[0008] The morphological risk assessment module is used to perform quantitative analysis on the medical image sequence of the target aneurysm through the morphological data processing channel to obtain a morphological risk assessment score that includes at least size, irregularity and growth trend components.

[0009] The fluid dynamics risk assessment module is used to calculate the spatiotemporal resolved blood flow data of the target aneurysm through the fluid dynamics data processing channel to obtain a fluid dynamics risk assessment score that includes at least the wall shear force distribution component, the oscillatory shear index component, and the blood flow impact component.

[0010] The spatiotemporal alignment module is used to spatiotemporally align the calculation results of the morphological data processing channel and the fluid dynamics data processing channel;

[0011] The weighted integration module is used to input the spatiotemporally aligned morphological risk assessment score and hydrodynamic risk assessment score into a preset weighted integration mechanism to generate an integrated risk score for the target aneurysm.

[0012] Preferably, the morphological risk assessment module performs quantitative analysis of the medical imaging sequences of the target aneurysm, including:

[0013] Extract the contour data of the target aneurysm from the medical image sequence;

[0014] Based on the contour data, the maximum size measurement, neck width measurement, and aspect ratio measurement of the target aneurysm are calculated, and the size sub-items are generated based on the maximum size measurement, neck width measurement, and aspect ratio measurement.

[0015] Based on the contour data, the surface curvature distribution of the target aneurysm is calculated, and the irregularity component is generated according to the statistical characteristics of the surface curvature distribution.

[0016] By comparing the medical image sequences at different time points, the differences in the contour data are identified, and the growth trend sub-item is generated based on the quantification results of the contour data differences.

[0017] Preferably, the hydrodynamic risk assessment module calculates the spatiotemporally resolved blood flow data of the target aneurysm by including:

[0018] Based on the contour data of the target aneurysm, the flow computation domain of the target aneurysm is constructed;

[0019] Within the flow computation domain, a preset set of control equations is solved to obtain the transient velocity field and pressure field inside the target aneurysm;

[0020] Based on the transient velocity field, the wall shear force vector field of the inner surface of the target aneurysm sac is calculated, and the spatial distribution of the wall shear force vector field is statistically analyzed to generate the wall shear force distribution components.

[0021] Based on the time-varying characteristics of the wall shear force vector field, the component of the oscillating shear index is calculated;

[0022] Based on the transient velocity field and pressure field, the blood flow kinetic energy distribution inside the target aneurysm is calculated, and combined with the spatial location of the blood flow impact area, the blood flow impact component is generated.

[0023] Preferably, the spatiotemporal alignment module performs spatiotemporal alignment of the computation results of the morphological data processing channel and the fluid dynamics data processing channel, including:

[0024] Establish a mapping relationship between the image coordinate system of the medical image sequence and the spatial coordinate system of the flow computation domain;

[0025] The contour data upon which the morphological risk assessment score depends is registered to the grid nodes of the flow computing domain through the mapping relationship;

[0026] The wall shear force vector field, on which the fluid dynamics risk assessment score depends, is mapped back to the image space of the medical image sequence through the inverse transformation of the mapping relationship;

[0027] Within the mapped image space, for each sampling point on the contour data, the wall shear force vector field data, oscillatory shear index component data, and blood flow impact component data corresponding to the spatial location are associated.

[0028] Preferably, the weighted integration mechanism includes:

[0029] Preset morphological and fluid dynamics fundamental weights;

[0030] Based on the values ​​of the growth trend components, the morphological basis weights are dynamically adjusted to obtain dynamic morphological weights.

[0031] Based on the values ​​of the oscillatory shear index components, the basic fluid dynamics weights are dynamically adjusted to obtain dynamic fluid dynamics weights;

[0032] The dynamic morphological weights and the dynamic fluid dynamics weights are normalized.

[0033] Using the normalized dynamic morphological weights, all sub-items in the morphological risk assessment score are weighted and fused to obtain a weighted morphological score.

[0034] Using the normalized dynamic fluid dynamics weights, all sub-items in the fluid dynamics risk assessment score are weighted and fused to obtain a weighted fluid dynamics score.

[0035] The integrated risk score is obtained by calculating the arithmetic mean of the weighted morphological score and the weighted hydrodynamic score.

[0036] Preferably, the dynamic adjustment of the morphological basis weights includes:

[0037] Set the growth trend influence coefficient;

[0038] Multiply the value of the growth trend component by the growth trend influence coefficient to obtain the weight adjustment amount;

[0039] The morphological basis weights are added to the weight adjustment amount to obtain the preliminary dynamic weights;

[0040] The initial dynamic weights are subjected to preset upper and lower limits to obtain the dynamic morphological weights.

[0041] Preferably, the dynamic adjustment of the fundamental fluid dynamics weights includes:

[0042] Set the oscillation shear effect coefficient;

[0043] Multiply the value of the oscillatory shear index component by the oscillatory shear influence coefficient to obtain the fluid dynamics weight adjustment amount;

[0044] Add the basic fluid dynamics weights to the fluid dynamics weight adjustment amount to obtain the preliminary fluid dynamics weights;

[0045] The preliminary fluid dynamic weights are subjected to preset upper and lower limits of fluid dynamic weights to obtain the dynamic fluid dynamic weights.

[0046] Preferably, after obtaining the integrated risk score, the method further includes:

[0047] Establish a correspondence between integrated risk scores and preset risk level labels;

[0048] Based on the specific value of the integrated risk score, a matching preset risk level label is found in the correspondence;

[0049] The found preset risk level labels are associated and encapsulated with the target aneurysm identification information, the integrated risk score, the weighted morphological score, and the weighted hydrodynamic score to form a structured risk assessment report.

[0050] Preferably, the process of generating the structured risk assessment report includes:

[0051] Create a report template that includes an identifier field, a rating field, a grade field, and a details field;

[0052] Fill the identification information of the target aneurysm into the identification field;

[0053] Enter the value of the integrated risk score into the scoring field;

[0054] Fill the preset risk level label into the level field;

[0055] The weighted morphological score, the weighted hydrodynamic score, the values ​​of each component of the morphological risk assessment score, and the values ​​of each component of the hydrodynamic risk assessment score are integrated and filled into the details field.

[0056] Preferably, the weighted integration mechanism further includes a verification step based on the consistency of individual items:

[0057] Compare the irregularity component in the morphological risk assessment score with the wall shear force distribution component in the fluid dynamics risk assessment score;

[0058] Calculate the correlation measure between the irregularity component and the wall shear force distribution component;

[0059] If the correlation metric is lower than a preset consistency threshold, a re-verification instruction is triggered for the segmentation result of the medical image sequence and the construction result of the flow computing domain.

[0060] Compared with the prior art, the beneficial effects of the present invention are:

[0061] Through the spatiotemporal alignment module, the geometric features output from the morphological data processing channel and the spatiotemporally resolved blood flow parameters output from the hydrodynamic data processing channel are precisely matched and correlated at the same spatial location and cardiac phase on the aneurysm wall. This allows for the direct acquisition of the specific magnitude, direction, and oscillation characteristics of the wall shear force experienced by the local area during systole or diastole in irregular regions such as aneurysm sacs and lobes. It overcomes the limitation of traditional methods that only allow for overall or statistical comparisons of the two types of parameters, achieving a leap from "isolated parameters" to "correlation maps." This enables intuitive visualization and quantitative analysis of the interaction between the hemodynamic environment and local morphological changes in the aneurysm, providing a fusion data foundation with clear spatiotemporal correspondences for a deeper understanding of the biological mechanisms of aneurysm growth, shaping, and potential rupture.

[0062] A pre-defined weighted integration mechanism is employed to automatically fuse and calculate the risk scores of multiple morphological and hydrodynamic components that have undergone spatiotemporal alignment. This mechanism does not simply add up fixed weights; instead, it learns and quantifies the contribution of complex synergistic or antagonistic relationships between different components to the final rupture risk through a nonlinear mapping function built based on prior knowledge or machine learning models. This integration approach replaces qualitative synthesis or simple linear scoring models that rely on clinical experience, generating a single, objective, repeatable, and more discriminative integrated risk score. This score not only integrates multi-dimensional information but also captures deep, nonlinear interaction patterns between morphological features and blood flow patterns. This allows for more accurate identification of complex cases where morphological indicators appear insignificant but blood flow patterns are highly dangerous, or vice versa, improving the robustness and predictive accuracy of the risk assessment system. Attached Figure Description

[0063] Figure 1 This is a time-series diagram of the aneurysm rupture risk prediction system based on medical imaging and hemodynamics described in this invention.

[0064] Figure 2 A flowchart for morphological quantitative analysis;

[0065] Figure 3 A flowchart for spatiotemporal alignment;

[0066] Figure 4 A bar chart showing the scores for the sub-items of the aneurysm morphological risk assessment.

[0067] Figure 5 A multi-indicator line-bar chart for assessing the risk of aneurysm rupture. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] Please see Figure 1This invention provides an aneurysm rupture risk prediction system based on medical imaging and hemodynamics. The system includes: a comprehensive risk assessment framework management module that initializes and coordinates the parallel or sequential execution of morphological data processing channels and fluid dynamics data processing channels. The morphological risk assessment module receives medical image sequences of the target aneurysm through the morphological data processing channel, performs quantitative analysis, and outputs a morphological risk assessment score including size, irregularity, and growth trend components. The fluid dynamics risk assessment module receives spatiotemporally resolved blood flow data of the target aneurysm through the fluid dynamics data processing channel, calculates, and outputs a fluid dynamics risk assessment score including wall shear force distribution, oscillatory shear index, and blood flow impact components. A spatiotemporal alignment module receives the calculation results from the two channels, establishes a spatial mapping relationship, and registers and correlates the morphological contour data with the fluid dynamics field data. A weighted integration module receives the spatiotemporally aligned scores, performs dynamic weight calculation and score fusion according to a preset weighted integration mechanism, and finally generates an integrated risk score for the target aneurysm.

[0070] Example 1: See Figure 2 The morphological risk assessment module's quantitative analysis process for medical image sequences is as follows: First, the contour data of the target aneurysm is extracted from the medical image sequence. Based on this contour data, the maximum size, neck width, and aspect ratio of the target aneurysm are calculated, and a size component is generated based on these measurements. Based on the same contour data, the surface curvature distribution of the target aneurysm is calculated, and an irregularity component is generated based on the statistical characteristics of this surface curvature distribution. Medical image sequences at different time points are compared to identify differences in the contour data, and a growth trend component is generated based on the quantitative results of these differences. The fluid dynamics risk assessment module's calculation process for spatiotemporally resolved blood flow data is as follows: Based on the aforementioned contour data of the target aneurysm, a flow computation domain for the target aneurysm is constructed. Within this flow computation domain, a preset set of governing equations is solved to obtain the transient velocity and pressure fields inside the target aneurysm. Based on this transient velocity field, the wall shear force vector field of the inner surface of the target aneurysm sac is calculated. The spatial distribution of this wall shear force vector field is statistically analyzed to generate a wall shear force distribution component. Based on the temporal variation characteristics of this wall shear force vector field, an oscillating shear index component is calculated. Based on the transient velocity field and pressure field, the blood flow kinetic energy distribution inside the target aneurysm is calculated, and the blood flow impact component is generated by combining the spatial location of the blood flow impact area.

[0071] In practical implementation, the morphological risk assessment module's quantitative analysis of medical image sequences involves specific data processing steps. The module extracts the contour data of the target aneurysm from a set of continuous medical image slices containing the aneurysm. This contour data consists of a series of three-dimensional spatial coordinate points, used to define the inner surface boundary of the target aneurysm sac wall. Based on the contour data, the morphological risk assessment module calculates multiple geometric parameters. In practice, the maximum size measurement calculated by the morphological risk assessment module is the maximum value of the spatial Euclidean distance between all point pairs in the contour data. The aneurysm neck width measurement is the narrowest distance at the opening connecting the aneurysm sac to the parent artery. The aspect ratio measurement is the ratio of the maximum aneurysm depth to the aneurysm neck width measurement. The size components are generated by linearly weighted summing these three measurements. For example, in a specific calculation, the maximum size measurement is 7.2 mm, the aneurysm neck width measurement is 4.1 mm, and the calculated aspect ratio measurement is 1.5. Based on contour data, the morphological risk assessment module quantifies irregularity by calculating the principal curvature of each point on the surface and analyzing its distribution characteristics. The irregularity component can be generated based on the statistical characteristics of the surface curvature distribution. A formula for quantifying the irregularity component is expressed as follows:

[0072]

[0073] in: This represents the values ​​of the irregularity component. This represents the standard deviation of the Gaussian curvature at all surface sampling points. This represents the mean of the absolute values ​​of the average curvature of all surface sampling points. and These are preset weighting coefficients, calculated from contour data in an exemplary calculation. It is 0.85. It is 0.32, in , Under the preset conditions, the calculated irregularity components The value is 0.691. The morphological risk assessment module evaluates the growth trend by comparing medical image sequences at different time points. In practice, the morphological risk assessment module loads two medical image sequences six months apart, extracts the contour data at the corresponding time points, and calculates their volume. The quantitative result of the difference in contour data can be expressed as the volume change rate. The growth trend sub-item is generated based on the volume change rate. For example, if the aneurysm volume measured in the first examination is 85 cubic millimeters, and the volume measured in the follow-up examination six months later is 92 cubic millimeters, the calculated volume growth rate is 8.2%. This quantitative result is directly mapped to the value of the growth trend sub-item.

[0074] In some embodiments, the fluid dynamics risk assessment module calculates spatiotemporally resolved blood flow data based on the contour data obtained from the aforementioned morphological analysis. The module constructs a flow computation domain for the target aneurysm based on the contour data and the geometric information of the proximal carrier artery. This flow computation domain is a three-dimensional volumetric mesh space used to define the computational region for blood flow simulation. Within this domain, the module solves the Navier-Stokes equations describing blood flow to obtain the transient velocity and pressure fields within the target aneurysm that change over time throughout the cardiac cycle. Based on the transient velocity field, the module calculates the wall shear force vector at each point on the inner surface of the target aneurysm sac wall. This wall shear force vector field is a spatial vector distribution, and the wall shear force distribution components are generated statistically from the spatial distribution of the wall shear force vector field, such as calculating the spatial average of the wall shear force magnitude, the percentage of areas below a specific threshold, and the spatial gradient. Based on the temporal variation characteristics of the wall shear force vector field, the module calculates the oscillatory shear index component at each point on the sac wall surface. This oscillatory shear index component quantifies the degree of oscillation in the direction of the wall shear force within a cardiac cycle. Based on the transient velocity and pressure fields, the fluid dynamics risk assessment module calculates the blood flow energy at each location inside the target aneurysm. The blood flow impact component is generated by identifying high-value areas of blood flow energy distribution and combining the spatial relationship between these areas and the aneurysm wall. For example, in a specific calculation, the area with the highest blood flow energy is identified as continuously impacting a local area at the top of the aneurysm. The value of the blood flow impact component reflects the relative size and impact intensity of the impact area.

[0075] Example 2: See Figure 3 A mapping relationship is established between the image coordinate system of the medical image sequence and the spatial coordinate system of the fluid computation domain. The contour data upon which the morphological risk assessment score depends is registered to the grid nodes of the fluid computation domain through this mapping relationship. Simultaneously, the wall shear force vector field upon which the fluid dynamics risk assessment score depends is mapped back to the image space of the medical image sequence through the inverse transformation of this mapping relationship. Within the mapped image space, each sampling point on the contour data is associated with the corresponding spatial location's wall shear force vector field data, oscillatory shear index component data, and blood flow impact component data.

[0076] In practical implementation, the spatiotemporal alignment module establishes a mapping relationship between the image coordinate system of the medical image sequence and the spatial coordinate system of the flow computation domain. The image coordinate system of the medical image sequence is defined by pixel indices and slice positions, while the spatial coordinate system of the flow computation domain is defined by physical length units (such as millimeters). The spatiotemporal alignment module calculates rigid body transformation or affine transformation matrices by identifying a set of corresponding feature points in the two coordinate systems. In a specific example, the pixel coordinates of three anatomical landmarks of the aneurysm neck in the medical image sequence are recorded in the image coordinate system, and the coordinates of the corresponding three spatial points in the flow computation domain are also recorded. The rotation matrix and translation vector are obtained by solving the least squares method, thereby defining the mapping relationship:

[0077]

[0078] in: Represents a physical space coordinate vector. Represents the image voxel coordinate vector. Represents the rotation matrix. This represents the translation vector.

[0079] In some embodiments, the establishment of the mapping relationship depends on the metadata of the medical image and the spatial scale parameters set during the construction of the mobile computing domain. The spatiotemporal alignment module registers the contour data upon which the morphological risk assessment score depends to the grid nodes of the mobile computing domain through the mapping relationship. The contour data consists of a series of three-dimensional points. The spatiotemporal alignment module applies a transformation formula to convert the image coordinates of each contour point to physical space coordinates, and finds the grid node in the volume or surface grid of the mobile computing domain that is closest to the spatial location of the physical coordinates, thus associating the contour point with the grid node. For example, a point in a set of contour points has coordinates (124, 87, 45) in the image coordinate system, and its physical coordinates are calculated to be (12.3, 5.1, 8.9) mm after mapping. The grid node coordinates (12.3, 5.0, 9.0) mm closest to this physical coordinate are found in the mobile computing domain, and the two are spatially associated.

[0080] It is understandable that the spatiotemporal alignment module simultaneously maps the wall shear force vector field, upon which the fluid dynamics risk assessment score depends, back to the image space of the medical image sequence through an inverse transformation of the mapping relationship. The wall shear force vector field data is originally defined at the spatial locations of grid nodes in the flow computation domain. The spatiotemporal alignment module inverses the mapping relationship, calculating the physical coordinates of each grid node back to the image voxel coordinates, and assigns the wall shear force vector of that node to the corresponding image spatial location. For locations in the image space that are not directly corresponding to grid nodes, the wall shear force vector value is determined through a spatial interpolation algorithm. Within the mapped image space, the spatiotemporal alignment module associates the wall shear force vector field data, oscillatory shear index component data, and blood flow impact component data for each sampling point on the contour data, corresponding to the spatial location. The spatiotemporal alignment module utilizes the established mapping and inverse mapping relationship between image space and physical space. For any sampling point on the contour data, its image coordinates are known. Using these coordinates, it directly queries or interpolates in the wall shear force vector field, oscillatory shear index sub-term data field, and blood flow impact sub-term data field mapped back to image space, thereby obtaining fluid dynamics data that completely corresponds to the spatial location of the sampling point. Optionally, the association process can construct a lookup table in the specific implementation. The table records the unique identifier of each contour sampling point, its coordinates in image space, the physical space coordinates after the mapping relationship transformation, the grid node number of the associated flow computation domain, and the wall shear force vector, oscillatory shear index sub-term value, and blood flow impact sub-term value obtained from the fluid dynamics field. The spatiotemporally aligned contour sampling point data will be used as the input of the weighted integration module.

[0081] Example 3: The weighted integration module presets morphological and fluid dynamics base weights. The morphological base weights are dynamically adjusted based on the growth trend component's value. Specifically, a growth trend influence coefficient is set, and the value of the growth trend component is multiplied by this coefficient to obtain the weight adjustment amount. The morphological base weight is then added to this adjustment amount to obtain the initial dynamic weight. Preset upper and lower limits are applied to the initial dynamic weight to obtain the dynamic morphological weight. The fluid dynamics base weights are dynamically adjusted based on the oscillation shear index component's value to obtain the dynamic fluid dynamics weight. The dynamic morphological and fluid dynamics weights are then normalized. The normalized dynamic morphological weights are used to weight and fuse all components of the morphological risk assessment score to obtain the weighted morphological score. The normalized dynamic fluid dynamics weights are used to weight and fuse all components of the fluid dynamics risk assessment score to obtain the weighted fluid dynamics score. The arithmetic mean of the weighted morphological score and the weighted fluid dynamics score is calculated to obtain the integrated risk score.

[0082] In practical implementation, the weighted integration module presets morphological and fluid dynamics base weights, which are initial static weighting factors used to balance the contributions of morphological and fluid dynamics information when dynamic factors are not considered. For example, the morphological base weight can be set to 0.5, and the fluid dynamics base weight can also be set to 0.5. The weighted integration module dynamically adjusts the morphological base weights based on the values ​​of the growth trend component. In practical implementation, the values ​​of the growth trend component are derived from the output of the morphological risk assessment module. A growth trend influence coefficient is set to quantify the influence of the growth trend on the morphological base weights. The weight adjustment amount is obtained by multiplying the value of the growth trend component by the growth trend influence coefficient. For example, if the value of the growth trend component is 8.2 and the growth trend influence coefficient is preset to 0.02, the calculated weight adjustment amount is 0.164. The initial dynamic weight is obtained by adding the morphological base weight and the weight adjustment amount. The dynamic morphological weight is obtained by applying the preset upper limit and lower limit constraints to the initial dynamic weight. For example, the initial dynamic weight is 0.664, which is obtained by adding the morphological base weight of 0.5 and the weight adjustment amount of 0.164. The upper limit constraint is set to 0.8 and the lower limit constraint is set to 0.3. Since 0.664 is in the interval [0.3, 0.8], the dynamic morphological weight is 0.664.

[0083] In some embodiments, the weighted integration module dynamically adjusts the basic fluid dynamics weights based on the values ​​of the oscillatory shear index component to obtain dynamic fluid dynamics weights. The values ​​of the oscillatory shear index component are derived from the output of the fluid dynamics risk assessment module. The weighted integration module normalizes the dynamic morphological weights and the dynamic fluid dynamics weights, ensuring that the sum of the two weights is 1. The calculation can be expressed as follows:

[0084]

[0085]

[0086] in: Represents dynamic morphological weights. Represents dynamic fluid dynamics weights. This represents the normalized dynamic morphological weights. This represents the normalized dynamic fluid dynamics weights. The normalized dynamic morphological weights are used to weight and fuse all components of the morphological risk assessment score to obtain a weighted morphological score. The morphological risk assessment score includes size, irregularity, and growth trend components. The weighted morphological score can be the sum of the products of each component value and a preset internal sub-weight, multiplied by the normalized dynamic morphological weights. Similarly, the normalized dynamic fluid dynamics weights are used to weight and fuse all components of the fluid dynamics risk assessment score to obtain a weighted fluid dynamics score. The fluid dynamics risk assessment score includes wall shear force distribution, oscillatory shear index, and blood flow impact components, and its fusion method is similar to that of the weighted morphological score. The weighted integration module calculates the arithmetic mean of the weighted morphological score and the weighted fluid dynamics score to obtain the integrated risk score. For example, if the weighted morphological score is 0.72 and the weighted fluid dynamics score is 0.85, the integrated risk score calculated by the arithmetic mean is 0.785. Optionally, when calculating the weighted morphological score and the weighted hydrodynamic score, the weighted integration module can use a linear weighting model to fuse the sub-items within each of them. The sub-item values ​​need to be standardized before fusion to eliminate the influence of dimensions.

[0087] See Figure 4 This is a bar chart showing the scores for the various sub-items in the aneurysm morphological risk assessment. The scores for each sub-item vary considerably among different patients. Most patients have moderate scores in the "Irregularity" sub-item, while the "Size" sub-item scores are relatively stable with no extremely low values. This chart is used to visually compare the sub-item performance of different patients in the morphological risk assessment, helping medical personnel quickly identify high-risk sub-items (such as patients with abnormal growth trends). It is one of the visualization outputs of the morphological risk assessment module in the aneurysm rupture risk prediction system.

[0088] Example 4: The specific process of dynamically adjusting the basic fluid dynamics weights in the weighted integration module is as follows: An oscillation shear influence coefficient is set; the value of the oscillation shear index component is multiplied by the oscillation shear influence coefficient to obtain the fluid dynamics weight adjustment amount; the basic fluid dynamics weight is added to this fluid dynamics weight adjustment amount to obtain the preliminary dynamic fluid dynamics weight; preset upper and lower limits of fluid dynamics weight constraints are applied to the preliminary dynamic fluid dynamics weight to obtain the dynamic fluid dynamics weight. After obtaining the integrated risk score, the system establishes a correspondence between the integrated risk score and preset risk level labels. Based on the specific value of the integrated risk score, a matching preset risk level label is searched in this correspondence. The found preset risk level label is associated and encapsulated with the target aneurysm identification information, integrated risk score, weighted morphological score, and weighted fluid dynamics score to form a structured risk assessment report.

[0089] In practical implementation, the process of dynamically adjusting the basic fluid dynamics weights in the weighted integration module is achieved through a preset oscillation shear influence coefficient. This coefficient is a scaling factor used to quantify the influence of the oscillation shear index component on the basic fluid dynamics weights. After setting the oscillation shear influence coefficient, the value of the oscillation shear index component is multiplied by the coefficient to obtain the fluid dynamics weight adjustment. For example, in a specific calculation scenario, if the oscillation shear index component value output by the fluid dynamics risk assessment module is 0.45, and the preset oscillation shear influence coefficient is 0.3, then the calculated fluid dynamics weight adjustment is 0.135. Adding the basic fluid dynamics weights to the adjusted weights yields the initial dynamic fluid dynamics weights. If the preset basic fluid dynamics weight is 0.5, then the initial dynamic fluid dynamics weight obtained after adding the values ​​is 0.635. Preset upper and lower limits for fluid dynamics weights are applied to the initial fluid dynamics weights to obtain the dynamic fluid dynamics weights. For example, if the upper limit constraint is set to 0.75 and the lower limit constraint is set to 0.25, and 0.635 is within the interval [0.25, 0.75], the dynamic fluid dynamics weight is determined to be 0.635. After obtaining the integrated risk score, the system establishes a correspondence between the integrated risk score and the preset risk level label. This correspondence can be a predefined mapping table, see Table 1.

[0090] Table 1: Correspondence between Integrated Risk Score and Risk Level

[0091]

[0092] The system searches for a matching preset risk level label in the corresponding relation based on the specific value of the integrated risk score. For example, if the integrated risk score is 0.785, Table 1 shows that it falls within the range of 0.7 to 1.0, therefore the matching preset risk level label is "high risk". In some embodiments, the matching process is implemented through program logic, traversing or binary searching a predefined score range to determine the corresponding level. The found preset risk level label is then associated and encapsulated with the target aneurysm's identification information, integrated risk score, weighted morphological score, and weighted hydrodynamic score to form a structured risk assessment report. This association and encapsulation organizes the scattered data items into a logical whole. Optionally, the identification information may include the patient number, examination date, and aneurysm location code; the weighted morphological score and weighted hydrodynamic score are intermediate results calculated by the weighted integration module. In an exemplary data encapsulation, the identification information is "P12345-20230915-ACom", the integrated risk score is 0.785, the preset risk level label is "high risk", the weighted morphological score is 0.72, and the weighted fluid dynamics score is 0.85. These data are collectively written into a structured data object or file. It can be understood that the structured risk assessment report provides a unified data format for subsequent display, storage, or transmission.

[0093] Example 5: The process of generating a structured risk assessment report involves creating a report template containing an identification field, a scoring field, a grading field, and a details field. The identification information of the target aneurysm is entered into the identification field, the integrated risk score value is entered into the scoring field, the preset risk grade label is entered into the grading field, and the weighted morphological score, weighted hydrodynamic score, the values ​​of each component of the morphological risk assessment score, and the values ​​of each component of the hydrodynamic risk assessment score are integrated and entered into the details field. The weighted integration mechanism also includes a verification step based on component consistency. This step compares the irregularity component in the morphological risk assessment score with the wall shear force distribution component in the hydrodynamic risk assessment score, calculates the correlation metric between the two, and if the correlation metric is lower than a preset consistency threshold, a re-verification instruction is triggered for the medical image sequence segmentation results and the flow computation domain construction results.

[0094] In practice, the generation process of a structured risk assessment report involves creating a report template that includes an identification field, a scoring field, a grading field, and a details field. The report template defines the data structure and the expected content format for each field. The identification information of the target aneurysm is entered into the identification field. This identification information is a string that uniquely identifies a risk assessment task; for example, "P12345-20230915-ACom" is entered into the identification field. The integrated risk score value is entered into the scoring field. This integrated risk score value is a floating-point number between 0 and 1; for example, an integrated risk score value of 0.785 is entered into the scoring field. The preset risk level label is entered into the grading field. This preset risk level label is a descriptive classification label; for example, "high risk" is entered into the grading field. The weighted morphological score, weighted hydrodynamic score, the individual component values ​​of the morphological risk assessment score, and the individual component values ​​of the hydrodynamic risk assessment score are integrated and entered into the details field. The integrated details field contains a data structure with multiple sub-items.

[0095] In some embodiments, the integration of detail fields is implemented using a specific data structure, such as a JavaScript object notation format, containing multiple subfields that store weighted morphological scores, weighted hydrodynamic scores, size component values, irregularity component values, growth trend component values, wall shear force distribution component values, oscillatory shear index component values, and blood flow impact component values, respectively. Optionally, the report template is defined in Extensible Markup Language (XML) format, and the final generated structured risk assessment report is saved as an XML file or database record. The weighted integration mechanism also includes a component consistency verification step, which compares the irregularity component in the morphological risk assessment score with the wall shear force distribution component in the hydrodynamic risk assessment score. In specific implementations, the comparison operation involves extracting the scalar or vector representation of the irregularity component from the calculated data and its corresponding representation with the wall shear force distribution component. For example, the irregularity component might be expressed as a feature vector based on curvature statistics, and the wall shear force distribution component might be expressed as a feature vector based on wall shear force spatial statistics. The correlation between the irregularity component and the wall shear force distribution component is calculated. The correlation can be calculated using the Pearson correlation coefficient, the formula of which is:

[0096]

[0097] in: The Pearson correlation coefficient represents the relationship between the irregularity component and the wall shear force distribution component. The first eigenvector representing the irregularity factor One element, The first eigenvector representing the partial eigenvectors of the wall shear force distribution One element, This represents the average of all elements in the irregularity feature vector. This represents the average value of all elements in the eigenvector of the wall shear force distribution. This represents the length of the eigenvector. In an exemplary calculation, the irregularity component eigenvector is [0.12, 0.25, 0.08], and the wall shear force distribution component eigenvector is [0.15, 0.30, 0.10]. The calculated Pearson correlation coefficient... The correlation coefficient is 0.998. If the correlation metric is lower than the preset consistency threshold, a re-verification instruction is triggered for the medical image sequence segmentation results and the mobile computational domain construction results. The consistency threshold can be set to 0.7. When the calculated Pearson correlation coefficient is lower than 0.7, the system generates and sends a re-verification instruction. The re-verification instruction can be an internal message containing a specific error code and problem description, notifying the system operator or triggering an automated script to re-execute the medical image sequence segmentation step and the mobile computational domain grid generation step.

[0098] See Figure 5 This is a multi-indicator line-bar chart for aneurysm rupture risk assessment. The weighted scores differ significantly among cases; the trend of the integrated risk score is generally consistent with the weighted morphological and hydrodynamic scores, reflecting the calculation logic of "averaging both." The integrated risk score of case P10008 is close to 0.35, which is relatively high in the current sample. This chart is a visualization of the "weighted integration module" in the aneurysm rupture risk prediction system, used to intuitively display the morphological and hydrodynamic risk contributions of different cases, as well as the final integrated risk level, helping medical personnel quickly identify high-risk cases and analyze the sources of risk.

[0099] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0100] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A system for predicting the risk of aneurysm rupture based on medical imaging and hemodynamics, characterized in that, The system includes: The integrated risk assessment framework management module is used to build and coordinate the morphological data processing channel and the fluid dynamics data processing channel; The morphological risk assessment module is used to perform quantitative analysis on the medical image sequence of the target aneurysm through the morphological data processing channel, extract the contour data of the target aneurysm from the medical image sequence, and obtain a morphological risk assessment score based on the contour data, which includes at least size, irregularity and growth trend components. The fluid dynamics risk assessment module is used to calculate the spatiotemporally resolved blood flow data of the target aneurysm through the fluid dynamics data processing channel, and obtain a fluid dynamics risk assessment score that includes at least a wall shear force distribution component, an oscillatory shear index component, and a blood flow impact component. The hydrodynamic risk assessment module performs calculations on the spatiotemporally resolved blood flow data of the target aneurysm, including: Based on the contour data of the target aneurysm, a flow computation domain of the target aneurysm is constructed. Within the flow computation domain, a set of preset control equations is solved to obtain the transient velocity field and pressure field inside the target aneurysm. Based on the transient velocity field, the wall shear force vector field of the inner surface of the target aneurysm sac is calculated, and the spatial distribution of the wall shear force vector field is statistically analyzed to generate the wall shear force distribution sub-items. The spatiotemporal alignment module is used to spatiotemporally align the calculation results of the morphological data processing channel and the fluid dynamics data processing channel; The weighted integration module is used to input the spatiotemporally aligned morphological risk assessment score and the hydrodynamic risk assessment score into a preset weighted integration mechanism to generate an integrated risk score for the target aneurysm. The spatiotemporal alignment module performs spatiotemporal alignment of the computation results of the morphological data processing channel and the fluid dynamics data processing channel, including: A mapping relationship is established between the image coordinate system of the medical image sequence and the spatial coordinate system of the flow computation domain; the contour data on which the morphological risk assessment score depends is registered to the grid nodes of the flow computation domain through the mapping relationship; the wall shear force vector field on which the hydrodynamic risk assessment score depends is mapped back to the image space of the medical image sequence through the inverse transformation of the mapping relationship; in the mapped image space, for each sampling point on the contour data, the wall shear force vector field data, oscillatory shear index component data, and blood flow impact component data corresponding to the spatial location are associated.

2. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 1, characterized in that, The morphological risk assessment module performs quantitative analysis of medical imaging sequences of the target aneurysm, including: Extract the contour data of the target aneurysm from the medical image sequence; Based on the contour data, the maximum size measurement, neck width measurement, and aspect ratio measurement of the target aneurysm are calculated, and the size sub-items are generated based on the maximum size measurement, neck width measurement, and aspect ratio measurement. Based on the contour data, the surface curvature distribution of the target aneurysm is calculated, and the irregularity component is generated according to the statistical characteristics of the surface curvature distribution. By comparing the medical image sequences at different time points, the differences in the contour data are identified, and the growth trend sub-item is generated based on the quantification results of the contour data differences.

3. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 2, characterized in that, The hydrodynamic risk assessment module also includes the following steps for calculating the spatiotemporally resolved blood flow data of the target aneurysm: Based on the time-varying characteristics of the wall shear force vector field, the component of the oscillating shear index is calculated; Based on the transient velocity field and pressure field, the blood flow kinetic energy distribution inside the target aneurysm is calculated, and combined with the spatial location of the blood flow impact area, the blood flow impact component is generated.

4. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 3, characterized in that, The weighted integration mechanism includes: Preset morphological and fluid dynamics fundamental weights; Based on the values ​​of the growth trend components, the morphological basis weights are dynamically adjusted to obtain dynamic morphological weights. Based on the values ​​of the oscillatory shear index components, the basic fluid dynamics weights are dynamically adjusted to obtain dynamic fluid dynamics weights; The dynamic morphological weights and the dynamic fluid dynamics weights are normalized. Using the normalized dynamic morphological weights, all sub-items in the morphological risk assessment score are weighted and fused to obtain a weighted morphological score. Using the normalized dynamic fluid dynamics weights, all sub-items in the fluid dynamics risk assessment score are weighted and fused to obtain a weighted fluid dynamics score. The integrated risk score is obtained by calculating the arithmetic mean of the weighted morphological score and the weighted hydrodynamic score.

5. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 4, characterized in that, The dynamic adjustment of the morphological basis weights includes: Set the growth trend influence coefficient; Multiply the value of the growth trend component by the growth trend influence coefficient to obtain the weight adjustment amount; The morphological basis weights are added to the weight adjustment amount to obtain the preliminary dynamic weights; The initial dynamic weights are subjected to preset upper and lower limits to obtain the dynamic morphological weights.

6. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 5, characterized in that, The dynamic adjustment of the fundamental fluid dynamics weights includes: Set the oscillation shear effect coefficient; Multiply the value of the oscillatory shear index component by the oscillatory shear influence coefficient to obtain the fluid dynamics weight adjustment amount; Add the basic fluid dynamics weights to the fluid dynamics weight adjustment amount to obtain the preliminary fluid dynamics weights; The preliminary fluid dynamic weights are subjected to preset upper and lower limits of fluid dynamic weights to obtain the dynamic fluid dynamic weights.

7. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 6, characterized in that, After obtaining the integrated risk score, the following is also included: Establish a correspondence between integrated risk scores and preset risk level labels; Based on the specific value of the integrated risk score, a matching preset risk level label is found in the correspondence; The found preset risk level labels are associated and encapsulated with the target aneurysm identification information, the integrated risk score, the weighted morphological score, and the weighted hydrodynamic score to form a structured risk assessment report.

8. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 7, characterized in that, The process of generating the structured risk assessment report includes: Create a report template that includes an identifier field, a rating field, a grade field, and a details field; Fill the identification information of the target aneurysm into the identification field; Enter the value of the integrated risk score into the scoring field; Fill the preset risk level label into the level field; The weighted morphological score, the weighted hydrodynamic score, the values ​​of each component of the morphological risk assessment score, and the values ​​of each component of the hydrodynamic risk assessment score are integrated and filled into the details field.

9. The aneurysm rupture risk prediction system based on medical imaging and hemodynamics according to claim 8, characterized in that, The weighted integration mechanism also includes a verification step based on the consistency of individual items: Compare the irregularity component in the morphological risk assessment score with the wall shear force distribution component in the fluid dynamics risk assessment score; Calculate the correlation measure between the irregularity component and the wall shear force distribution component; If the correlation metric is lower than a preset consistency threshold, a re-verification instruction is triggered for the segmentation result of the medical image sequence and the construction result of the flow computing domain.

Images