Method for predicting occlusion after intracranial unruptured aneurysm blood flow guiding treatment

Abstract

The invention provides a method for predicting occlusion after intracranial unruptured aneurysm blood flow guiding treatment. The method comprises the following steps: acquiring clinical data and DSA imaging data of a patient receiving treatment of a blood flow guiding device; the DSA iconography data comprises DSA data of preoperative, postoperative immediate and follow-up visit; based on DSA data immediately before and after an operation, aneurysm morphological parameters and hemodynamic parameters immediately before and after the operation are measured through three-dimensional reconstruction, finite element simulation and computational fluid mechanics methods; on the basis of follow-up DSA data, clinical data, aneurysm morphological parameters and hemodynamic parameters, a statistical analysis prediction model is adopted, and the occlusion condition after intracranial unruptured aneurysm blood flow guiding treatment is predicted. According to the scheme, intracranial aneurysm occlusion after FD treatment can be accurately predicted.

Description

Technical Field

[0001] The present invention belongs to the field of medical imaging and artificial intelligence technology, and particularly relates to a method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm. Background Art

[0002] Aneurysm refers to an abnormal protrusion of the arterial wall formed by acute or chronic dilation of the intracranial artery. It is a cerebrovascular disease that seriously threatens human life and health. Many studies have found that the prevalence of unruptured intracranial aneurysm (UIA) in the normal population is 2%-7%, and its development and rupture lead to subarachnoid haemorrhage (SAH). Statistics show that the incidence of SAH worldwide is about 9.1 / 100,000, and the mortality rate is as high as 25-50%; and more than half of the survivors suffer from lifelong neurological dysfunction and have a low overall quality of life. Therefore, precise intervention and treatment measures should be taken for patients with intracranial aneurysms.

[0003] At present, in addition to regular imaging follow-up, surgery is an important treatment for UIA, mainly including craniotomy clipping and endovascular intervention. In the past two decades, with the invention and advent of interventional materials and corresponding interventional technologies such as embolic coils, auxiliary balloons, intracranial stents, and flow diverters (FD), coupled with the fact that endovascular interventional treatment is less invasive and has a fast recovery, it has now become the preferred surgical method for the treatment of UIA. FD is a new type of blood flow remodeling device with high metal coverage and high mesh density. Its treatment principle is different from the previous simple coil embolization, balloon-assisted embolization or stent combined embolization, and is aimed at reconstructing the blood flow of the aneurysm and the parent artery. Studies have shown that the placement of FD at the neck of the aneurysm can change the direction of blood flow from the parent vessel to the aneurysm, reduce the blood flow impact on the aneurysm wall, improve the hemodynamics in the aneurysm sac, thereby promoting the process of intratumoral thrombosis and ultimately achieving aneurysm healing. At the same time, FD implantation provides a "scaffold" for the creeping growth of new endothelial cells. Over time, vascular endothelial cells gradually cover the stent grid, achieve endothelialization of the aneurysm neck, permanently close the aneurysm neck, isolate the aneurysm cavity from the tumor-bearing blood vessel, achieve lasting occlusion of the aneurysm, and significantly reduce the risk of rupture. FD was first used in large / giant wide-necked aneurysms and showed excellent efficacy and safety. Since then, due to its unique and advanced treatment concept, its clinical application has gradually expanded to small and medium-sized aneurysms, dissecting / blood bubble aneurysms, distal aneurysms of the Willis circle, and vertebral basilar artery aneurysms. A meta-analysis of FD treatment of UIA showed that the long-term occlusion rate of aneurysms over 1 year can reach 77%-96%, while delayed rupture and ischemic stroke of aneurysms are rare and almost no serious delayed side effects are caused.

[0004] Although FD has excellent performance in clinical treatment, as its indications expand, some aneurysms are still incompletely occluded after FD implantation, with continuous visualization of the aneurysm and a high risk of rupture; therefore, exploring the key related factors of UIA occlusion after blood flow-guided treatment is a cutting-edge and hot issue in aneurysm research. Many studies have explored the impact of some clinical factors on incomplete aneurysm occlusion, and the results suggest that patient age, smoking history, aneurysm diameter, fusiform aneurysm, and follow-up time may be independent risk factors for incomplete aneurysm occlusion.

[0005] The treatment concept of FD is to achieve aneurysm occlusion by improving the hemodynamic status of the aneurysm, and abnormal hemodynamic status is considered to be the key mechanism for the formation, enlargement, rupture and recurrence of intracranial aneurysms. The hemodynamic analysis of intracranial aneurysms based on computational fluid dynamics (CFD) can obtain morphological parameters, evaluate the hemodynamic changes of aneurysms before or after surgery, and predict the risk of rupture. The CFD method provides precise visual evidence for exploring the effect of hemodynamics on aneurysm occlusion after FD implantation. Studies have shown that the morphology of the tumor-bearing vessel changes after FD implantation, and the increase in the blood flow incidence angle (the angle between the incident blood flow and the aneurysm top) promotes the healing of the aneurysm. Many studies have also confirmed that the reduction of aneurysm blood flow inflow, incident blood flow velocity and wall shear stress (WSS) after FD treatment is closely related to the occlusion of the aneurysm.

[0006] This application aims to develop a prediction model for complete occlusion of UIA after FD treatment, accurately predict the time-dependent occlusion of UIA after FD treatment, help doctors make reasonable individualized treatment decisions, and reduce the risk of bleeding. Summary of the invention

[0007] The purpose of the embodiments of this specification is to provide a method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm.

[0008] To solve the above technical problems, the embodiments of the present application are implemented in the following ways:

[0009] In a first aspect, the present application provides a method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm, the method comprising:

[0010] Clinical data and DSA imaging data of patients who received blood flow diversion device treatment were obtained; DSA imaging data included preoperative, immediate postoperative and follow-up DSA data;

[0011] Based on the preoperative and immediate postoperative DSA data, the aneurysm morphological parameters and preoperative and immediate postoperative hemodynamic parameters were determined by three-dimensional reconstruction, finite element simulation, and computational fluid dynamics methods;

[0012] Based on the follow-up DSA data, clinical information, aneurysm morphological parameters and hemodynamic parameters, a statistical analysis prediction model was used to predict the occlusion status of unruptured intracranial aneurysms after blood flow-directed treatment.

[0013] In one embodiment, based on the DSA data before and immediately after the operation, the aneurysm morphological parameters and the hemodynamic parameters before and immediately after the operation are determined by three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods, including:

[0014] Performing three-dimensional reconstruction on the preoperative DSA data to obtain a first geometric model of the aneurysm-parent artery;

[0015] Determining the morphological parameters according to the first geometric model;

[0016] Performing three-dimensional reconstruction on the immediate postoperative DSA data to obtain a second geometric model of the aneurysm-parent artery;

[0017] According to the parameters of the blood flow guiding device, the FD three-dimensional model is determined;

[0018] Determine the three-dimensional model of the spring coil according to the parameters of the spring coil;

[0019] The preoperative and immediate postoperative hemodynamic parameters are determined based on the first geometric model, the second geometric model, the FD three-dimensional model and the spring coil three-dimensional model using finite element simulation and fluid mechanics analysis methods.

[0020] In one embodiment, the selection of variables in the statistical analysis prediction model includes:

[0021] According to the follow-up DSA data, the occlusion status of the aneurysm was determined; the cases were divided into occlusion group and non-occlusion group;

[0022] Obtaining candidate variables; wherein the candidate variables include the patient's clinical data, aneurysm morphological parameters, preoperative hemodynamic parameters, and changes in postoperative hemodynamic parameters;

[0023] Univariate analysis was performed for each candidate variable between the aneurysm occlusion group and the non-occlusion group, and the statistical difference of each candidate variable was evaluated according to the P value;

[0024] The Lasso regression method was used to select variables related to aneurysm occlusion;

[0025] According to the results of univariate analysis and the relevant variables selected by Lasso regression method, the variables in the statistical analysis prediction model were determined.

[0026] In one embodiment, the construction of the statistical analysis prediction model includes:

[0027] Several machine learning algorithms were performed on the variables selected in the statistical analysis prediction model to develop a prediction model for complete UIA occlusion after FD treatment;

[0028] Several evaluation methods are used for internal evaluation of model performance;

[0029] The prediction model obtained by the machine learning algorithm corresponding to the highest AUC value in the evaluation results is used as the statistical analysis prediction model.

[0030] In one embodiment, the several machine learning algorithms include multivariate logistic regression, random forest, Xgboost and KNN.

[0031] In one embodiment, the plurality of evaluation methods include: repeated 5-fold cross validation, 0.632 bootstrap, and 0.632+bootstrap.

[0032] In one embodiment, SHAP is used to quantify the contribution of each variable to the prediction of the statistical analysis prediction model;

[0033] Factors associated with UIA occlusion after FD treatment were determined based on the corresponding contribution of each variable.

[0034] In a second aspect, the present application provides a device for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm, the device comprising:

[0035] An acquisition module is used to acquire clinical data and DSA imaging data of patients treated with blood flow diversion devices; the DSA imaging data include preoperative, immediate postoperative and follow-up DSA data;

[0036] A measurement module is used to measure aneurysm morphological parameters and hemodynamic parameters before and after surgery based on preoperative and immediate postoperative DSA data through three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods;

[0037] The prediction module is used to predict the occlusion status of unruptured intracranial aneurysms after blood flow-guided treatment based on follow-up DSA data, clinical information, aneurysm morphological parameters and hemodynamic parameters, using a statistical analysis prediction model.

[0038] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm according to the first aspect is implemented.

[0039] In a fourth aspect, the present application provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm as described in the first aspect.

[0040] It can be seen from the technical solution provided in the above embodiments of this specification that this solution can accurately predict the time-dependent UIA occlusion status after FD treatment, which is of great significance for doctors to make reasonable individualized treatment decisions and reduce the risk of bleeding. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0042] Figure 1 A schematic diagram of the process of predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm provided in this application;

[0043] Figure 2 Summary graph of SHAP values ​​provided for this application;

[0044] Figure 3 The purpose of this application is to provide;

[0045] Figure 4 The probability of aneurysm occlusion predicted by five predictive factors and the SHAP analysis diagram of the predictive analysis provided in this application;

[0046] Figure 5 A schematic diagram of the structure of the device for predicting occlusion after blood flow-guided treatment of unruptured intracranial aneurysms provided in this application;

[0047] Figure 6 A schematic diagram of the structure of the electronic device provided in this application. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0049] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0050] It will be apparent to those skilled in the art that various modifications and variations may be made to the specific embodiments of the present application description without departing from the scope or spirit of the present application. Other embodiments derived from the present application description will be apparent to those skilled in the art. The present application description and examples are exemplary only.

[0051] The words “include,” “including,” “have,” “contain,” etc. used in this document are open-ended terms, meaning including but not limited to.

[0052] Unless otherwise specified, "parts" in this application are all calculated by mass.

[0053] The present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0054] Reference Figure 1 , which shows a flow chart of a method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm provided in an embodiment of the present application.

[0055] like Figure 1 As shown in the figure, the methods for predicting occlusion after flow-guided treatment of unruptured intracranial aneurysms may include:

[0056] S110. Obtain clinical data and DSA imaging data of patients treated with blood flow diversion devices; DSA imaging data include preoperative, immediate postoperative and follow-up DSA data. Optionally, clinical data may include: baseline data, clinical manifestations, treatment status, aneurysm distribution location, and aneurysm morphology.

[0057] Specifically, the baseline data, clinical manifestations, treatment status, aneurysm distribution, aneurysm morphology, DSA imaging data and follow-up results of patients with different types of aneurysms treated with flow diverter (FD) in the neurosurgery department of each hospital were collected. The baseline data mainly included: patient age, gender, past medical history (mainly history of hypertension and diabetes), personal history (mainly history of smoking and drinking), family history, etc. Clinical manifestations were divided into headache, dizziness, limb symptoms, other and asymptomatic. Treatment status mainly included the type of FD, whether it was combined with coil embolization, and the results of immediate postoperative angiography. The main sites of aneurysms were the internal carotid artery, middle cerebral artery, vertebral basilar artery, etc. The morphology of aneurysms can be divided into regular aneurysms and irregular aneurysms (irregular type: defined as lobed or with daughter sacs). The types of aneurysms can be divided into saccular aneurysms and dissecting / blood bubble-like aneurysms. Whether the aneurysm has perforating arteries (yes: defined as perforating arteries originating from any position of the aneurysm wall, except for those originating from the junction of the aneurysm-bearing artery and the aneurysm wall). DSA imaging data include preoperative, immediate postoperative and follow-up DSA data of patients. Follow-up data include follow-up time, DSA images, etc.

[0058] S120. Based on the preoperative and immediate postoperative DSA data, the aneurysm morphological parameters and preoperative and immediate postoperative hemodynamic parameters are calculated by three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods, which may include:

[0059] Performing three-dimensional reconstruction on the preoperative DSA data to obtain a first geometric model of the aneurysm-parent artery;

[0060] Determining the morphological parameters according to the first geometric model;

[0061] Performing three-dimensional reconstruction on the immediate postoperative DSA data to obtain a second geometric model of the aneurysm-parent artery;

[0062] According to the parameters of the blood flow guiding device, the FD three-dimensional model is determined;

[0063] Determine the three-dimensional model of the spring coil according to the parameters of the spring coil;

[0064] The hemodynamic parameters are determined according to the first geometric model, the second geometric model, the FD three-dimensional model and the spring coil three-dimensional model using finite element simulation and fluid mechanics analysis methods.

[0065] Specifically, the level set segmentation method can be used to perform three-dimensional reconstruction on the preoperative and immediate postoperative 3D-DSA data to obtain the true geometric model of the patient's aneurysm-parent artery, and to intercept the aneurysm-parent artery and the surrounding related main branches or perforators for subsequent analysis and calculation.

[0066] It can be understood that the geometric model is for each patient, and when predicting the aneurysm occlusion of the patient, three-dimensional reconstruction is performed based on the DSA data of the patient to obtain the corresponding geometric model.

[0067] It can also be understood that based on the first geometric model, the FD three-dimensional model and the coil three-dimensional model, the preoperative hemodynamic parameters can be determined using finite element simulation and fluid dynamics analysis methods; based on the second geometric model, the FD three-dimensional model and the coil three-dimensional model, the immediate postoperative hemodynamic parameters can be determined using finite element simulation and fluid dynamics analysis methods.

[0068] Based on the virtual treatment method of intracranial aneurysms, the finite element method (FEM) was used to complete the complete virtual implantation process of FD and coils.

[0069] Among them, the virtual implantation process of FD includes:

[0070] (1) Reconstruct the FD three-dimensional model in a free state based on the geometric characteristics and material information of the selected blood flow guidance device.

[0071] (2) The reconstructed FD three-dimensional model is delivered along the centerline of the blood vessel to the target position on the geometric model of the aneurysm-parent artery, simulating the actual surgical process.

[0072] (3) The reconstructed FD three-dimensional model is released to fit the blood vessel wall through its own stress-strain field to obtain a three-dimensional model after release, that is, a FD three-dimensional model.

[0073] The parameters of the spring coil include the model of the spring coil.

[0074] The virtual coil implantation process includes:

[0075] (1) Generate the center line of the spring coil according to the spring coil model;

[0076] (2) The coils are pushed into the aneurysm one by one, and the centerline of the coils bends and twists according to the geometric shape of the aneurysm cavity based on its mechanical properties;

[0077] (3) Converting the center line of the spring coil after release into a three-dimensional model of the spring coil. The release refers to the process of bending and twisting along with the geometric shape of the aneurysm cavity, and the center line of the spring coil after release is the center line of the spring coil that is bent and twisted along with the geometric shape of the aneurysm cavity.

[0078] Generally, during the implantation of FD and coil, the arterial wall is set as a rigid body by default. The released FD / coil 3D model is exported for subsequent hemodynamic simulation analysis.

[0079] The preoperative and postoperative models were divided into three-dimensional grids, and hemodynamic simulation was performed. The preoperative model includes the first geometric model of the aneurysm-parent artery, and the postoperative model includes the second geometric model of the aneurysm-parent artery and the three-dimensional model of the FD / coil. Due to the large differences in the geometric dimensions of the blood vessels, FD and coils, their grid sizes were also determined to be different values. The grid size of the FD was set to 0.02mm, and the grid size of the blood vessels was set to 0.25mm. The blood was set to a laminar, incompressible Newtonian fluid with a density of 1066kg / m3 and a viscosity of 0.0035kg / m·s; the vascular wall was set to a rigid, non-slip boundary. The inlet flow rate setting refers to the average vascular flow rate reported in previous studies. The inlet flow rate of the vertebral artery was set to 1.3ml / s, and the inlet flow rate of the internal carotid artery (ICA) was set to 4.6ml / s. The outlet conditions were calculated according to Murray's law:

[0080]

[0081] Among them, Q i is the flow rate at the outlet, D i is the diameter of the outlet, and n is 3. According to the law of conservation of mass, the sum of the outlet flow rates is equal to the sum of the inlet flow rates, and the flow rates of each outlet can be obtained.

[0082] The above models and boundary conditions are used in combination, and the Navier-Stokes equation is applied to perform steady-state simulation calculations. The hemodynamic parameters are calculated based on the simulation results, and the blood flow characteristics in the aneurysm are displayed using visualization methods such as three-dimensional streamlines and velocity isosurfaces. Among them, the outlet condition is one of the boundary conditions.

[0083] Among them, the hemodynamic parameters include: WSS (wall shear stress), NWSS (normalized wall shear stress), blood flow inflow, relative inflow, aneurysm average flow velocity, parent artery average flow velocity, standardized velocity, RFV (residual flow volume) and ICI, etc. Among them, WSS refers to the friction force generated by blood flow on the blood vessel wall per unit area; NWSS refers to the ratio of the aneurysm WSS and the parent artery WSS; blood flow inflow refers to the blood flow that flows into the aneurysm through the aneurysm neck per unit time; relative inflow refers to the ratio of blood flow inflow to the parent artery blood flow; the ratio of the average blood flow velocity in the aneurysm and the average blood flow velocity in the parent artery is defined as the standardized velocity; RFV refers to the volume of the velocity area greater than a certain set velocity value. Since the flow velocity of the parent artery in different cases is different, the velocity threshold used to reflect the change of the high-velocity area in the aneurysm before and after surgery is different; ICI describes the concentration of blood flow entering the aneurysm, which refers to the ratio of the proportion of the flow in the parent artery into the aneurysm to the proportion of the area of ​​the inflow area above the aneurysm neck:

[0084]

[0085] Among them, Q in is the flow rate into the aneurysm, Q v is the flow rate of the parent artery, A in A is the area of ​​the inflow region of the tumor neck, o The area of ​​the tumor neck.

[0086] New morphological parameters OsR (ostium ratio) and NR (neck ratio) are used to quantify the extent of aneurysm encapsulation of the parent artery. The calculation formula is as follows:

[0087]

[0088] Among them, A ostium is the surface area of ​​the aneurysm neck, A vessel is the surface area of ​​the non-lesioned area of ​​the aneurysm segment on the parent artery. Similarly, NR is the parameter of OsR based on two-dimensional images, and the calculation formula is:

[0089]

[0090] Among them, CND is the width of the aneurysm neck, D1 and D2 are the diameters of the parent artery at the proximal and distal ends of the aneurysm neck, respectively. vessel It may be 0, so we propose an improved OsR:

[0091]

[0092] On this basis, two other morphological parameters were defined to quantify the degree of aneurysm expansion, namely ArR (area ratio) and VoR (volume ratio). ArR is defined as the ratio of the aneurysm surface area to the reconstructed healthy vessel surface area; VoR is defined as the ratio of the aneurysm volume to the reconstructed healthy vessel volume.

[0093] S130. Based on the follow-up DSA data, clinical information, aneurysm morphological parameters and hemodynamic parameters, a statistical analysis prediction model was used to predict the occlusion status of unruptured intracranial aneurysms after flow-directed treatment.

[0094] Specifically, according to the follow-up DSA, the cases were divided into occlusion group and non-occlusion group, and the clinical indicators, morphological parameters and hemodynamic parameters between the groups were analyzed by univariate and multivariate analysis to establish the UIAs occlusion prediction model (i.e., statistical analysis prediction model) and to identify the factors related to UIA occlusion after FD treatment. The prediction performance of the training optimization model was evaluated by Repeated 5-fold cross-validation, 0.632 Bootstrap, and 0.632 + Bootstrap. Finally, Shapley additive explanation analysis (SHAP analysis) was used to determine the contribution of each variable to the prediction of unruptured aneurysm occlusion.

[0095] Among them, categorical variables (i.e., count data) in the statistical analysis prediction model were expressed as number (rate). Continuous variables (i.e., measurement data) were expressed as mean ± standard deviation (mean ± SD). Two-sided tests were adopted, and significant statistical differences were observed when P < 0.05. Among them, independent sample t-test was used for inter-group comparison of measurement data. If the inter-group measurement data did not meet the t-test conditions, the Wilcoxon rank sum test was used. The chi-square test, continuity-corrected chi-square test, or Fisher's exact probability method were used for inter-group comparison of count data.

[0096] In one embodiment, the selection of variables in the statistical analysis prediction model includes the following:

[0097] Based on the follow-up DSA data, the occlusion status of the aneurysm was determined, and the cases were divided into an occluded group (O'Kelly–Marotta (OKM) D grade) and a non-occluded group (OKM AC grade);

[0098] Obtaining candidate variables; wherein the candidate variables include the patient's clinical data, aneurysm morphological parameters, preoperative hemodynamic parameters, and changes in postoperative hemodynamic parameters;

[0099] Univariate analysis was performed for each candidate variable between the aneurysm occlusion group and the non-occlusion group, and the statistical difference of each candidate variable was evaluated according to the P value;

[0100] The Lasso regression method was used to select variables related to aneurysm occlusion;

[0101] According to the results of univariate analysis and the relevant variables selected by Lasso regression method, the variables in the statistical analysis prediction model were determined.

[0102] Specifically, the patients' clinical data, preoperative parameters (morphological parameters, hemodynamic parameters), and changes in postoperative hemodynamic parameters were considered as candidate variables. Univariate analysis was performed on each predictor variable, and their statistical differences were evaluated based on the P value. Lasso regression (a machine learning method) was then used to automatically select variables related to aneurysm occlusion. Neurosurgeons reviewed the important variables in the univariate analysis and Lasso regression and ultimately determined the final variables to be included in the model. Based on the above variable selection steps, the five selected variables were follow-up time, OsR, change in WSS (Pa) of the parent artery, ArR, and WSS (Pa) of the parent artery.

[0103] In one embodiment, the construction of the statistical analysis prediction model includes:

[0104] Several machine learning algorithms were performed on the variables selected in the statistical analysis prediction model to develop a prediction model for complete UIA occlusion after FD treatment;

[0105] Use multiple evaluation methods to conduct internal evaluation of model performance;

[0106] The prediction model obtained by the machine learning algorithm corresponding to the highest AUC (Area Under Curve) value in the evaluation results is used as the statistical analysis prediction model.

[0107] Among them, several machine learning algorithms include multivariate logistic regression, random forest, Xgboost and KNN (K nearest neighbor); several evaluation methods include: repeated 5-fold cross validation, 0.632 bootstrap, 0.632+ bootstrap. It can be understood that the machine learning algorithm and evaluation method can also use other algorithms or methods that can meet the functional requirements, and there is no limitation here.

[0108] Specifically, we performed machine learning algorithms such as multivariate logistic regression, random forest, Xgboost, and KNN on the selected variables to develop a prediction model for predicting complete occlusion of unruptured intracranial aneurysms after blood flow diversion device treatment. The model performance was internally evaluated using three methods: Repeated 5-fold cross-validation, 0.632 Bootstrap, and 0.632 + Bootstrap. After the evaluation, the model AUC value obtained by multivariate logistic regression analysis was the highest (AUC is defined as the area under the ROC (Receiver Operating Characteristic) curve and the coordinate axis. Obviously, the value of this area will not be greater than 1. Since the ROC curve is generally above the straight line y = x, the value range of AUC is between 0.5 and 1. The closer the AUC is to 1.0, the higher the authenticity of the detection method.) Therefore, the prediction model obtained by multivariate logistic regression analysis was selected for subsequent analysis.

[0109] In one embodiment, SHAP analysis is used to quantify the contribution of each variable to the prediction of the statistical analysis prediction model;

[0110] According to the contribution value corresponding to each variable, the factors related to UIA occlusion after FD treatment were determined.

[0111] Specifically, the SHAP value is based on the Shapley value in game theory, which assigns a "fair" importance score by considering all possible feature combinations. For a machine learning model, the Shapley value can be used to measure the contribution of each feature to the prediction result. Specifically, the Shapley value calculates the average marginal contribution of a feature to the model output under all possible feature combinations.

[0112] Figure 2 The figure is a summary plot of SHAP values, where Change of Parent Artery WSS is the change value of WSS (Pa) of the parent artery, Parent Artery WSS is the WSS of the parent artery, and Follow-up Time Value is the follow-up time. The figure is sorted according to the average value of the absolute SHAP value of each feature. The length of each bar in the figure represents the average absolute SHAP value of the feature in all samples in the data set, reflecting its importance. The longer the bar, the greater the importance. The length of the bar intuitively shows the importance of each feature on the model prediction. As shown in the figure, the change of the parent artery wall shear stress (WSS) is the most important feature that contributes most to accurate prediction.

[0113] In particular, force plots can be generated to understand the specific situation of each patient. Figure 3 ), the bold 0.13 represents the model's output score for this patient's outcome. A higher score makes the model tend to predict an outcome of 1 (UIA occlusion), while a lower score encourages the model to predict an outcome of 0 (UIA non-occlusion). The variables in the model are highlighted in red and blue. Red indicates variables that improve the model's score, while blue indicates variables that decrease the score. Variables that have a greater impact on the score are located closer to the red and blue dividing lines (e.g., feature 5 and feature 0), and the size of the bar indicates the degree of influence of the variable.

[0114] For each patient, the probability of aneurysm occlusion predicted by the five predictive factors, namely, follow-up time, OsR, change in WSS (Pa) of the parent artery, ArR and WSS (Pa) of the parent artery, and the SHAP analysis diagram of the prediction analysis were generated (see Figure 2). Figure 4 ) to comprehensively consider the factors that affect the aneurysm occlusion after blood flow diversion therapy and the possibility of aneurysm occlusion, so as to timely adjust the postoperative medication regimen and select the appropriate DSA angiography follow-up time. For example, Figure 4 When the follow-up time is 2, OsR is 0.117, the change value of WSS (Pa) of the parent artery is -5.695, ArR is 0.335, and WSS (Pa) of the parent artery is 8.113, the predicted probability of aneurysm occlusion is 0.7437, and the corresponding SHAP analysis diagram of the predictive analysis is shown. From the SHAP analysis diagram, it can be seen that the factor of WSS (Pa) change value of the parent artery is more important.

[0115] The method for predicting occlusion after blood flow-guided treatment of unruptured intracranial aneurysms provided in the embodiments of the present application can accurately predict the occlusion of time-dependent UIA after FD treatment, which is of great significance for doctors to make reasonable individualized treatment decisions and reduce the risk of bleeding.

[0116] Reference Figure 5 , which shows a schematic structural diagram of an apparatus for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm described according to one embodiment of the present application.

[0117] like Figure 5 As shown, the device 500 for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm may include:

[0118] An acquisition module 510 is used to acquire clinical data and DSA imaging data of patients treated with a blood flow guiding device; the DSA imaging data includes preoperative, immediate postoperative and follow-up DSA data;

[0119] The determination module 520 is used to determine the aneurysm morphological parameters and the hemodynamic parameters before and after the operation based on the DSA data before and after the operation through three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods;

[0120] The prediction module 530 is used to predict the occlusion status of unruptured intracranial aneurysms after blood flow-guided treatment by using a statistical analysis prediction model based on the follow-up DSA data, clinical information, aneurysm morphological parameters and hemodynamic parameters.

[0121] Optionally, the determination module 520 is further configured to:

[0122] Performing three-dimensional reconstruction on the preoperative DSA data to obtain a first geometric model of the aneurysm-parent artery;

[0123] Determining the morphological parameters according to the first geometric model;

[0124] Performing three-dimensional reconstruction on the immediate postoperative DSA data to obtain a second geometric model of the aneurysm-parent artery;

[0125] According to the parameters of the blood flow guiding device, the FD three-dimensional model is determined;

[0126] Determine the three-dimensional model of the spring coil according to the parameters of the spring coil;

[0127] The hemodynamic parameters are determined according to the first geometric model, the second geometric model, the FD three-dimensional model and the spring coil three-dimensional model using finite element simulation and fluid mechanics analysis methods.

[0128] Optionally, the device further includes: a variable selection module, configured to:

[0129] According to the follow-up DSA data, the occlusion of the aneurysm was determined and the cases were divided into occlusion group and non-occlusion group.

[0130] Obtaining candidate variables; wherein the candidate variables include the patient's clinical data, aneurysm morphological parameters, preoperative hemodynamic parameters, and changes in postoperative parameters;

[0131] Univariate analysis was performed for each candidate variable between the aneurysm occlusion group and the non-occlusion group, and the statistical difference of each candidate variable was evaluated according to the P value;

[0132] The Lasso regression method was used to select variables related to aneurysm occlusion;

[0133] According to the results of univariate analysis and the relevant variables selected by Lasso regression method, the variables in the statistical analysis prediction model were determined.

[0134] Optionally, the device further comprises a model building module, for:

[0135] Several machine learning algorithms were performed on the variables selected in the statistical analysis prediction model to develop a prediction model for complete UIA occlusion after FD treatment;

[0136] Several evaluation methods are used for internal evaluation of model performance;

[0137] The prediction model obtained by the machine learning algorithm corresponding to the highest AUC value in the evaluation results is used as the statistical analysis prediction model.

[0138] Optionally, several machine learning algorithms including Multivariate Logistic Regression, Random Forest, Xgboost, and KNN.

[0139] Optionally, several evaluation methods include: Repeated 5-fold cross-validation, 0.632Bootstrap, 0.632+Bootstrap.

[0140] Optionally, SHAP is used to quantify the contribution of each variable to the prediction of the statistical analysis prediction model;

[0141] Factors associated with UIA occlusion after FD treatment were determined based on the corresponding contribution of each variable.

[0142] The present embodiment provides a device for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm, which can execute the above method embodiments. Its implementation principles and technical effects are similar and will not be described in detail here.

[0143] Figure 6 FIG. 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 6 As shown, a structural schematic diagram of an electronic device 600 suitable for implementing an embodiment of the present application is shown.

[0144] like Figure 6 As shown, the electronic device 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 are also stored. The CPU 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0145] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage section 608 as needed.

[0146] In particular, according to the embodiments of the present disclosure, the above reference Figure 1 The described process can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes program code for executing the above-mentioned intracranial unruptured aneurysm blood flow guided treatment after occlusion prediction method. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611.

[0147] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of the code, and the aforementioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0148] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not constitute limitations on the units or modules themselves in certain circumstances.

[0149] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop, a mobile phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0150] As another aspect, the present application further provides a storage medium, which may be the storage medium included in the aforementioned device in the above embodiment; or may be a storage medium that exists independently and is not assembled into the device. The storage medium stores one or more programs, and the aforementioned programs are used by one or more processors to execute the method for predicting occlusion after blood flow-guided treatment of unruptured intracranial aneurysms described in the present application.

[0151] Storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0152] It should be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0153] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

Claims

1. A method for predicting occlusion after blood flow-guided treatment of unruptured intracranial aneurysms, characterized in that: The method comprises: Obtain clinical data and DSA imaging data of patients who received blood flow diversion device treatment; the DSA imaging data include preoperative, immediate postoperative and follow-up DSA data; Based on the preoperative and immediate postoperative DSA data, the aneurysm morphological parameters and the preoperative and immediate postoperative hemodynamic parameters are determined by three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods; Based on the follow-up DSA data, clinical information, aneurysm morphological parameters and hemodynamic parameters, a statistical analysis prediction model is used to predict the occlusion status of unruptured intracranial aneurysms after blood flow-guided treatment.

2. The method according to claim 1, characterized in that The aneurysm morphological parameters and the hemodynamic parameters before and after the operation are determined based on the DSA data before and after the operation by three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods, including: Performing three-dimensional reconstruction on the preoperative DSA data to obtain a first geometric model of the aneurysm-parent artery; Determining the morphological parameters according to the first geometric model; Performing three-dimensional reconstruction on the immediate postoperative DSA data to obtain a second geometric model of the aneurysm-parent artery; According to the parameters of the blood flow guiding device, the FD three-dimensional model is determined; Determine the three-dimensional model of the spring coil according to the parameters of the spring coil; The preoperative and immediate postoperative hemodynamic parameters are determined based on the first geometric model, the second geometric model, the FD three-dimensional model and the spring coil three-dimensional model using finite element simulation and fluid dynamics analysis methods.

3. The method according to claim 1, characterized in that The selection of variables in the statistical analysis prediction model includes: According to the follow-up DSA data, the occlusion status of the aneurysm was determined, and the cases were divided into occluded group and non-occluded group; Obtaining candidate variables; wherein the candidate variables include the patient's clinical data, aneurysm morphological parameters, preoperative hemodynamic parameters, and changes in postoperative hemodynamic parameters; Perform univariate analysis on each of the candidate variables between the aneurysm occlusion group and the non-occlusion group, and evaluate the statistical difference of each candidate variable according to the P value; The Lasso regression method was used to select variables related to aneurysm occlusion; According to the univariate analysis results and the relevant variables selected by the Lasso regression method, the variables in the statistical analysis prediction model are determined.

4. The method according to claim 3, characterized in that The construction of the statistical analysis prediction model includes: executing several machine learning algorithms on the selected variables in the statistical analysis prediction model to develop a prediction model for complete UIA occlusion after FD treatment; Several evaluation methods are used for internal evaluation of model performance; The prediction model obtained by the machine learning algorithm corresponding to the highest AUC value in the evaluation results is used as the statistical analysis prediction model.

5. The method according to claim 4, characterized in that The several machine learning algorithms include multivariate logistic regression, random forest, Xgboost and KNN.

6. The method according to claim 4, characterized in that The several evaluation methods include: repeated 5-fold cross validation method, 0.632 bootstrap method, and 0.632+bootstrap method.

7. The method according to any one of claims 3 to 6, characterized in that: SHAP is used to quantify the contribution of each of the variables to the prediction of the statistical analysis prediction model; Factors associated with UIA occlusion after FD treatment were determined based on the corresponding contribution of each of the above variables.

8. A device for predicting occlusion after blood flow-guided treatment of unruptured intracranial aneurysms, characterized in that: The device comprises: An acquisition module, used to acquire clinical data and DSA imaging data of patients treated with a blood flow guiding device; the DSA imaging data includes preoperative, immediate postoperative and follow-up DSA data; A determination module, for determining aneurysm morphological parameters and hemodynamic parameters before and after the operation based on the DSA data before and after the operation by three-dimensional reconstruction, finite element simulation and computational fluid dynamics methods; The prediction module is used to predict the occlusion status of unruptured intracranial aneurysm after blood flow-guided treatment by using a statistical analysis prediction model based on the follow-up DSA data, clinical information, aneurysm morphological parameters and the hemodynamic parameters.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm as described in any one of claims 1-7 is implemented.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for predicting occlusion after blood flow-guided treatment of an unruptured intracranial aneurysm as described in any one of claims 1 to 7 is implemented.

Images