Methods, devices, apparatuses, and media for prognosis prediction of tissue hypoperfusion after thrombectomy

Abstract

The embodiment of the present disclosure provides a method and device for predicting the prognosis of post-stent retrieval tissue hypoperfusion, equipment and medium, which relates to the technical field of medical intelligent analysis, and includes: collecting clinical feature information of a patient, preoperative multi-modal brain images before stent retrieval and postoperative multi-modal brain images after stent retrieval; determining preoperative perfusion feature parameters based on the preoperative multi-modal brain images, determining postoperative perfusion feature parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multi-modal brain images; constructing a multi-dimensional feature vector by using the clinical feature information, the preoperative perfusion feature parameters, the postoperative perfusion feature parameters and the postoperative tissue hypoperfusion phenotype information, inputting the multi-dimensional feature vector into a prognosis prediction model, and obtaining the prediction state information of the prognosis of the patient output by the prognosis prediction model. The present application can improve the prediction accuracy of prognosis.

Description

Technical Field

[0001] This disclosure relates to the field of medical intelligent analysis technology, and in particular to a method, device, equipment and medium for predicting the prognosis of tissue hypoperfusion after thrombectomy. Background Technology

[0002] In patients with acute ischemic stroke, cerebral blood vessels are blocked by thrombi, leading to cerebral ischemia and hypoxia. Mechanical thrombectomy uses interventional devices to directly remove or break up the thrombus blocking the blood vessel, rapidly restoring cerebral blood flow. After undergoing endovascular mechanical thrombectomy, nearly half of the patients achieve good vascular recanalization but still have a poor prognosis.

[0003] Currently, prognostic prediction for patients with acute ischemic stroke after mechanical thrombectomy mainly relies on clinical scoring systems, imaging indicators, and statistical models. Among these, methods based on clinical scoring systems depend on the surgeon's experience and visual interpretation, are easily influenced by subjectivity, and exhibit significant inter-observer variability; methods based on imaging indicators and statistical models are based on only single-modality data, resulting in low predictive accuracy. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems, or at least partially solve the above-mentioned technical problems, this disclosure provides a method, device, equipment and medium for predicting the prognosis of tissue hypoperfusion after thrombectomy.

[0005] This disclosure provides a method for predicting the prognosis of tissue hypoperfusion after thrombectomy. The method includes: collecting the patient's clinical characteristics, preoperative multimodal brain images before thrombectomy, and postoperative multimodal brain images after thrombectomy; determining preoperative perfusion characteristic parameters based on the preoperative multimodal brain images, and determining postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multimodal brain images; constructing a multidimensional feature vector using the clinical characteristics, preoperative perfusion characteristic parameters, postoperative perfusion characteristic parameters, and postoperative tissue hypoperfusion phenotype information; inputting the multidimensional feature vector into a prognosis prediction model to obtain the predicted prognosis status information of the patient output by the prognosis prediction model; wherein, the prognosis prediction model is trained using multidimensional feature vector samples of historical patients and corresponding prognostic results, and the predicted prognosis status information of the patient is an indicator used to characterize the development of tissue hypoperfusion after surgery.

[0006] In this implementation, the application comprehensively considers the patient's clinical characteristics, pre- and post-thrombectomy multimodal imaging information to solve the problems of single data source and heterogeneity, thereby improving prediction accuracy. It quantitatively assesses vascular recanalization and tissue perfusion status through pre- and post-operative perfusion characteristic parameters, assesses post-operative perfusion risk through post-operative tissue hypoperfusion phenotype information, and assesses the patient's basic physical condition through clinical characteristics. By combining the three dimensions of characteristic data, it achieves automated prediction of the patient's prognosis, which can guide early intervention or adjuvant therapy, ultimately improving the patient's functional recovery and reducing mortality.

[0007] In one possible implementation, preoperative perfusion characteristic parameters are determined based on preoperative multimodal brain images, including: dividing the preoperative multimodal brain images into infarct and non-infarct areas based on the signal intensity of pixels in the preoperative multimodal brain images; determining the preoperative perfusion characteristic parameters of the overall region based on the time-varying pattern of contrast agent concentration in the infarct and non-infarct areas of the preoperative multimodal brain images; and determining the preoperative perfusion characteristic parameters of the target region based on the time-varying pattern of contrast agent concentration in the infarct area of ​​the preoperative multimodal brain images; wherein the preoperative perfusion characteristic parameters include the mean maximum delay time, the mean cerebral blood flow, and the mean cerebral blood volume.

[0008] In one possible implementation, postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information are determined based on postoperative multimodal brain images, including: dividing the postoperative multimodal brain images into infarct and non-infarct areas based on the signal intensity of pixels in the postoperative multimodal brain images; determining postoperative perfusion characteristic parameters of the overall region based on the time-varying pattern of contrast agent concentration in the infarct and non-infarct areas of the postoperative multimodal brain images; and determining postoperative perfusion characteristic parameters of the target region based on the time-varying pattern of contrast agent concentration in the infarct area of ​​the postoperative multimodal brain images; wherein, the postoperative perfusion characteristic parameters include the mean maximum delay time, the mean cerebral blood flow, and the mean cerebral blood volume.

[0009] In this implementation, the infarct area and non-infarct area are objectively divided by the intensity of the image signal, which can solve the problem of subjective variation in personal experience and visual judgment. Furthermore, by collecting a variety of perfusion characteristic parameters before and after the operation, the problem of single data source and heterogeneity can be solved, thereby improving the prediction accuracy, generalization and real-time performance. This helps clinicians identify high-risk patients, guide early intervention or adjuvant treatment, and ultimately improve patients' functional recovery and reduce mortality.

[0010] In one possible implementation, postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information are determined based on postoperative multimodal brain images, including: determining relative cerebral blood flow based on the infarct area and determining the maximum delay time based on the non-infarct area; combining the relative cerebral blood flow and the maximum delay time to obtain postoperative tissue hypoperfusion phenotype information in the postoperative multimodal brain images; wherein, the postoperative tissue hypoperfusion phenotype information includes tissue hypoperfusion type and tissue hypoperfusion volume.

[0011] In this implementation, the application specifically determines the relative cerebral blood flow in the infarct area and the maximum delay time in the non-infarct area, thereby matching the perfusion parameters with the low perfusion mechanisms in different regions. By combining the relative cerebral blood flow and the maximum delay time for accurate identification of multimodal brain images, it can effectively distinguish the low perfusion types in the infarct area and the non-infarct area. This solves the problem that the traditional eTICI score cannot distinguish low perfusion of different mechanisms and has large inter-observer variability, providing a standardized, mechanism-specific assessment tool.

[0012] In one possible implementation, the collection of patients' clinical characteristic information includes: collecting multidimensional characteristic information of historical patients based on preset multidimensional features, and collecting postoperative status development information of historical patients; wherein, the multidimensional features include various basic clinical features, disease-related features, NIHSS scores, and recanalization levels; performing correlation analysis on the multidimensional feature information and the corresponding postoperative status development information, and calculating the correlation coefficient; selecting target features from the multidimensional features whose correlation coefficients meet the correlation conditions to construct clinical features; and collecting patients' clinical characteristic information based on the clinical features.

[0013] In one possible implementation, preoperative multimodal brain imaging includes CT angiography and CT perfusion imaging, while postoperative multimodal brain imaging includes CT perfusion imaging and magnetic resonance diffusion-weighted imaging.

[0014] In one possible implementation, the patient's prognostic status information includes predictive functional status information after a preset time period post-surgery, predictive mortality risk information, and predictive complication status information.

[0015] This disclosure also provides a prognostic prediction device for tissue hypoperfusion after thrombectomy. The device includes: a data acquisition module for acquiring the patient's clinical characteristics, preoperative multimodal brain images, and postoperative multimodal brain images; an analysis module for determining preoperative perfusion characteristic parameters based on the preoperative multimodal brain images, and determining postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multimodal brain images; and a prediction module for constructing a multidimensional feature vector using the clinical characteristics, preoperative perfusion characteristic parameters, postoperative perfusion characteristic parameters, and postoperative tissue hypoperfusion phenotype information, inputting the multidimensional feature vector into a prognostic prediction model, and obtaining the patient's prognostic prediction status information output by the prognostic prediction model. The prognostic prediction model is trained using multidimensional feature vector samples from historical patients and corresponding prognostic results, and the patient's prognostic prediction status information is an indicator used to characterize the development of tissue hypoperfusion after surgery.

[0016] This disclosure also provides a computing device, which includes: a processor; a memory for storing processor-executable instructions; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the prognostic prediction method for tissue hypoperfusion after thrombectomy as provided in this disclosure.

[0017] This disclosure also provides a computer-readable storage medium storing a computer program for performing a prognostic method for post-thrombectomy tissue hypoperfusion as provided in this disclosure. Attached Figure Description

[0018] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0019] Figure 1 A flowchart illustrating a method for predicting the prognosis of tissue hypoperfusion after thrombectomy, provided in an embodiment of this disclosure; Figure 2 This is a schematic diagram of the prognostic device for post-thrombectomy tissue hypoperfusion provided in an embodiment of the present disclosure; Figure 3 This is a schematic diagram of the structure of a computing device provided in an embodiment of the present disclosure. Detailed Implementation

[0020] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0021] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0022] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0023] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0024] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0025] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0026] Prognostic prediction for patients with acute ischemic stroke after mechanical thrombectomy mainly relies on clinical scoring systems, imaging indicators, and statistical models. However, these methods do not incorporate brain tissue-level blood perfusion status, i.e., the phenotype of brain hypoperfusion, leading to an inability to distinguish different hypoperfusion mechanisms and thus affecting predictive accuracy. The eTICI assessment system relies on operator experience and visual interpretation, is susceptible to subjective influence, and exhibits significant inter-observer variability; furthermore, this system cannot distinguish hypoperfusion phenotypes. Multi-model approaches, based on only single-modality data and failing to integrate multimodal imaging and clinical features, lead to predictive bias and pose a risk of overfitting; while internal validation yields high accuracy, external validation performance deteriorates.

[0027] To address the aforementioned issues, this disclosure provides a method for predicting the prognosis of tissue hypoperfusion after thrombectomy. The method includes: collecting the patient's clinical characteristics, preoperative multimodal brain images, and postoperative multimodal brain images; determining preoperative perfusion characteristic parameters based on the preoperative multimodal brain images, and determining postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multimodal brain images; constructing a multidimensional feature vector using the clinical characteristics, preoperative perfusion characteristic parameters, postoperative perfusion characteristic parameters, and postoperative tissue hypoperfusion phenotype information; inputting the multidimensional feature vector into a prognosis prediction model to obtain the predicted prognosis status information of the patient output by the prognosis prediction model; wherein the prognosis prediction model is trained using multidimensional feature vector samples of historical patients and corresponding prognostic results, and the predicted prognosis status information of the patient is an indicator used to characterize the development of tissue hypoperfusion after surgery. This application comprehensively considers the patient's clinical characteristics, pre- and post-thrombectomy multimodal imaging information, and addresses the issues of single data source and heterogeneity, thereby improving prediction accuracy. It quantitatively assesses vascular recanalization and tissue perfusion status through pre- and post-operative perfusion characteristic parameters, evaluates post-operative perfusion risk through post-operative tissue hypoperfusion phenotype information, and assesses the patient's baseline physical condition through clinical characteristic information. By combining these three dimensions of characteristic data, it achieves automated prediction of the patient's prognosis, guides early intervention or adjuvant therapy, and ultimately improves the patient's functional recovery and reduces mortality.

[0028] The method will be described below with reference to specific embodiments.

[0029] Figure 1 This is a flowchart illustrating a prognostic prediction method for tissue hypoperfusion after thrombectomy, provided in an embodiment of this disclosure. This method can be executed by a prognostic prediction method for tissue hypoperfusion after thrombectomy, wherein the device can be implemented using software and / or hardware, and is generally integrated into a computing device. Figure 1 As shown, the method includes: S101. Collect the patient's clinical characteristics, preoperative multimodal brain images before thrombectomy, and postoperative multimodal brain images after thrombectomy.

[0030] Preoperative brain imaging was acquired before the patient underwent endovascular thrombectomy (EVT). Postoperative brain imaging was acquired within a predetermined time period after the endovascular thrombectomy. For example, the predetermined time period was within 48 hours postoperatively.

[0031] Multimodal brain imaging includes images that are routinely available in clinical settings. In one possible implementation, preoperative multimodal brain imaging includes Magnetic Resonance Diffusion-Weighted Imaging (DWI) and Computed Tomography Perfusion Imaging (CTP), while postoperative multimodal brain imaging includes CT perfusion imaging and magnetic resonance diffusion-weighted imaging.

[0032] Specifically, CT perfusion imaging and magnetic resonance diffusion-weighted imaging were acquired using standard clinical equipment and according to the Digital Imaging and Communications in Medicine (DICOM) standard within 48 hours before and after the patient's EVT procedure. Standard clinical equipment included Siemens and GE CT / MRI scanners. The CT perfusion imaging sequence consisted of dynamic scans with a scan time greater than 30 seconds, covering the entire brain, and an image resolution greater than 1 mm × 1 mm × 1 mm. For magnetic resonance diffusion-weighted imaging, the b-value was set to 1000 s / mm. 2 The image resolution is greater than 0.5mm × 0.5mm × 3mm. Furthermore, the acquired preoperative and postoperative CT perfusion imaging and magnetic resonance diffusion-weighted imaging are stored in the PACS system.

[0033] Clinical characteristic information of historical patients is collected in advance based on preset multidimensional features. The clinical characteristic information is determined according to the correlation between the multidimensional characteristic information of historical patients and the postoperative condition development information.

[0034] Understandably, among the various clinical characteristics of patients, there are those that are poorly correlated with or unrelated to the development of postoperative tissue hypoperfusion. Therefore, it is necessary to screen the multidimensional characteristics and select those that can affect the development of postoperative tissue hypoperfusion in order to construct clinical characteristic information.

[0035] Specifically, based on preset multidimensional features, historical patients' multidimensional feature information and the corresponding postoperative status development information are collected.

[0036] In one possible implementation, the multidimensional characteristic information of historical patients includes various basic clinical characteristics, disease-related characteristics, NIHSS scores, recanalization grades, etc.

[0037] For example, basic clinical characteristics include age, sex, past medical history, smoking history, baseline blood pressure, and baseline blood glucose levels. Onset-related characteristics include time to onset, time from onset to puncture, and occurrence of symptomatic intracerebral hemorrhage.

[0038] In one possible implementation, postoperative status development information includes the patient's prognosis 90 days after surgery, such as good prognosis, risk of death, and complications of cerebral hemorrhage.

[0039] Furthermore, the multidimensional feature information is quantified. For numerical features, the original values ​​are directly used as the quantification results, such as age, baseline blood pressure, baseline blood glucose level, onset time, time from onset to puncture, NIHSS score, and recanalization grade. For categorical features, a coding method is used for quantification, such as gender, past medical history, and smoking history. For example, male is assigned a value of 1, female is assigned a value of 0, past medical history is assigned a value of 1, no past medical history is assigned a value of 0, smoking history is assigned a value of 1, and no smoking history is assigned a value of 0, etc.

[0040] The postoperative outcome information is quantified. For example, a good prognosis is assigned a value of 1, the risk of death is assigned a value of 2, and the transformation of cerebral hemorrhage into complications is assigned a value of 3.

[0041] Statistical correlation analysis was used to analyze the correlation between the collected multidimensional feature information and the corresponding postoperative tissue hypoperfusion status development results, and the correlation coefficient was calculated. Target features with correlation coefficients meeting the correlation criteria were selected from the multidimensional features to construct the clinical features. The statistical correlation analysis methods could include Pearson correlation coefficient, Spearman's rank correlation coefficient, etc.

[0042] In one possible implementation, the Pearson correlation coefficient method is used to calculate the correlation between each multidimensional feature and the postoperative condition development outcome information, obtaining the Pearson correlation coefficient. The Pearson correlation coefficient ranges from -1 to 1; a larger absolute value indicates a stronger correlation between the feature and the postoperative condition development outcome information, while a smaller absolute value indicates a weaker correlation. Further, features with correlation coefficients greater than a preset value are selected as target features. For example, features with correlation coefficients greater than 0.6 are selected as target features to construct clinical features.

[0043] Based on the clinical characteristics identified above, collect the corresponding clinical characteristic information of the patients.

[0044] For example, if the above method is used to determine clinical characteristics including NIHSS score, smoking history, time from onset to puncture, occurrence of symptomatic intracerebral hemorrhage, and recanalization grade, then the corresponding clinical characteristic information of the patient is collected from the PACS system and electronic medical records. The collected clinical characteristic information is a structured table.

[0045] Further, the collected clinical feature information is preprocessed. Optionally, data cleaning is performed by removing missing values ​​and filling missing data with the median, and the clinical feature data is normalized using the z-score normalization method. Preprocessing is also performed on the magnetic resonance diffusion-weighted imaging. Optionally, noise reduction, grayscale normalization, and artifact correction are applied to the magnetic resonance diffusion-weighted imaging.

[0046] S102. Determine preoperative perfusion characteristic parameters based on preoperative multimodal brain imaging, and determine postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on postoperative multimodal brain imaging.

[0047] First, preoperative and postoperative multimodal brain images were processed using the same method to determine preoperative and postoperative perfusion characteristic parameters. Specific steps included: In one possible implementation, multimodal brain images are divided according to the signal intensity of pixels in the multimodal brain images to obtain infarct areas and non-infarct areas.

[0048] Specifically, magnetic resonance diffusion-weighted imaging is divided based on the signal intensity of pixels in magnetic resonance diffusion-weighted imaging to obtain infarct areas and non-infarct areas in magnetic resonance diffusion-weighted imaging.

[0049] In one possible implementation, the method for dividing the infarct area into the non-infarct area is the threshold method.

[0050] The signal intensity of each pixel in the standardized magnetic resonance diffusion-weighted imaging is extracted. The signal intensity of each pixel is compared with the signal intensity threshold. The region with a signal intensity greater than or equal to the signal intensity threshold is a high signal region, which is divided into an infarct region. The region with a signal intensity less than the signal intensity threshold is a low signal region, which is divided into a non-infarct region.

[0051] In one possible implementation, the infarct and non-infarct regions are divided using a deep learning model. Specifically, a semantic segmentation model is used to perform pixel-level automatic segmentation of magnetic resonance diffusion-weighted imaging. For example, the semantic segmentation model could be UNet, FCN, the DeepLab series, or SegNet. This model is pre-trained using manually labeled magnetic resonance diffusion-weighted imaging samples and includes a connected input layer, encoder, bottleneck layer, decoder, and output layer, which are used for feature downsampling, feature fusion, feature uptesting, and pixel-level classification of the magnetic resonance diffusion-weighted imaging, respectively.

[0052] Specifically, for the defined infarct and non-infarct regions, the number of pixels within the infarct region is calculated to obtain the infarct region volume. The infarct region volume is the product of the number of pixels in the infarct region, the pixel length spacing, the pixel width spacing, and the scan layer thickness. The pixel spacing and layer thickness parameters are extracted from DICOM image metadata.

[0053] In this implementation, by comparing the pixel signal intensity thresholds of magnetic resonance diffusion-weighted imaging, the infarct area and the non-infarct area can be objectively and accurately identified, avoiding the bias of traditional subjective interpretation.

[0054] Furthermore, the region segmentation results of magnetic resonance diffusion-weighted imaging are mapped onto CT perfusion imaging.

[0055] Specifically, the image registration method is used to align CT perfusion imaging and magnetic resonance diffusion-weighted imaging. The infarct area in magnetic resonance diffusion-weighted imaging is mapped to CT perfusion imaging to obtain the infarct area in CT perfusion imaging, and the non-infarct area in magnetic resonance diffusion-weighted imaging is mapped to CT perfusion imaging to obtain the non-infarct area in CT perfusion imaging.

[0056] More specifically, a strategy combining rigid and flexible registration was employed. Rigid registration was performed on CT perfusion imaging and MRI diffusion-weighted imaging, using the skull anatomy as a reference to correct spatial positional discrepancies between the two images. Flexible registration was then performed based on the grayscale distribution characteristics of brain tissue to compensate for local deviations caused by slight shifts in scanning position and tissue. After registration, the corresponding infarct and non-infarct areas were simultaneously determined in CT perfusion imaging based on the infarct and non-infarct boundaries defined in the MRI diffusion-weighted imaging.

[0057] Understandably, this step outputs a binary mask map of CT perfusion imaging and marks the boundaries between infarcted and non-infarcted areas.

[0058] The above methods were used to classify preoperative multimodal brain images and postoperative multimodal brain images.

[0059] Furthermore, perfusion characteristic parameters were calculated for both preoperative and postoperative multimodal brain images, targeting both the overall and target regions. The overall region included both infarcted and non-infarcted areas, while the target region included the infarcted area.

[0060] Specifically, dynamic curves of contrast agent concentration versus time were extracted for both the overall region and the target region. A deconvolution algorithm was then used to calculate the mean maximum delay time, mean cerebral blood flow, and mean cerebral blood volume. The mean maximum delay time (Tmax) was used to determine delayed perfusion. The mean cerebral blood volume (CBV) is the blood volume per unit mass of brain tissue, reflecting the microvascular perfusion reserve of the brain. The mean cerebral blood flow (CBF) is the volume of blood flowing through a unit mass of brain tissue per unit time, reflecting the actual blood perfusion rate of the brain.

[0061] Using the methods described above, the perfusion characteristic parameters of the overall region and the perfusion characteristic parameters of the target region corresponding to the preoperative multimodal brain images and postoperative multimodal brain images were calculated respectively.

[0062] In this implementation, the infarct area and non-infarct area are objectively divided by the intensity of the image signal, which can solve the problem of subjective variation in personal experience and visual judgment. Furthermore, by collecting a variety of perfusion characteristic parameters before and after the operation, the problem of single data source and heterogeneity can be solved, thereby improving the prediction accuracy, generalization and real-time performance. This helps clinicians identify high-risk patients, guide early intervention or adjuvant treatment, and ultimately improve patients' functional recovery and reduce mortality.

[0063] Furthermore, postoperative multimodal brain imaging was used to further determine the phenotypic information of postoperative tissue hypoperfusion.

[0064] Specifically, the relative cerebral blood flow is determined in the infarcted area, and the maximum delay time is determined based on the non-infarcted area.

[0065] In one possible implementation, relative cerebral blood flow is determined based on the infarct area from CT perfusion imaging.

[0066] Specifically, using the longitudinal fissure of the brain as the axis of symmetry, a region with the same volume and shape as the infarct area is selected at the corresponding position in the healthy hemisphere to obtain the target area. Further, CT perfusion imaging data is processed, and the blood flow signal is inverted using the arterial input function (AIF) and venous output function (VOF) to calculate the first cerebral blood flow corresponding to each pixel in the infarct area and the second cerebral blood flow corresponding to each pixel in the target area. For each pixel in the infarct area, the ratio of the first cerebral blood flow in the infarct area to the second cerebral blood flow corresponding to the mirror pixel in the target area is calculated to obtain the relative cerebral blood flow of that pixel. Specifically, the relative cerebral blood flow is: rCBF=CBF1 / CBF2 In the formula, rCBF is the relative cerebral blood flow, CBF1 is the first cerebral blood flow, and CBF2 is the second cerebral blood flow.

[0067] Furthermore, the open-source software RAPID was used to statistically analyze the distribution of relative cerebral blood flow within the infarct area, and to calculate the mean, standard deviation, and percentiles.

[0068] In this implementation, the healthy side mirror region is used as a reference to calculate the relative cerebral blood flow for the infarct area. This can quantify the degree of perfusion loss and effectively eliminate the interference of confounding factors such as individual baseline blood flow and scanning parameters, providing an objective data basis for assessing the perfusion status of the infarct area.

[0069] In one possible implementation, the maximum delay time is determined based on the non-infarcted area of ​​CT perfusion imaging.

[0070] Specifically, CT perfusion imaging data is processed, and a time-density curve is constructed based on the change of contrast agent concentration over time. The time when the contrast agent concentration in the non-infarct area reaches its peak, i.e., the maximum delay time T, is then calculated using the time-density curve. max .

[0071] Furthermore, the presence of a low-perfusion zone can be determined within the non-infarct area based on the maximum delay time.

[0072] Specifically, a maximum delay time threshold is determined. Generally, the maximum delay time threshold is 6 seconds. The maximum delay time is compared with the maximum delay time threshold. When the maximum delay time is greater than the maximum delay time threshold, it is determined that low perfusion may exist, and the area where low perfusion may exist is classified as a low perfusion zone.

[0073] Furthermore, regional morphological identification was performed on low-irrigation areas.

[0074] Specifically, the low-perfusion area is divided into multiple grid regions using a gridding method. For each grid region, the eccentricity and the number of connected components are identified at the pixel level. When the eccentricity of the low-perfusion area is greater than the eccentricity threshold and the number of connected components is less than the number threshold, the region is determined to have a wedge-shaped distribution along the blood vessel. For example, the eccentricity threshold is 0.7 and the number threshold is 5.

[0075] In this implementation, low-perfusion areas are first screened for non-infarcted areas using the maximum delay time threshold. Then, quantitative indicators such as eccentricity and number of connected components are used to accurately determine whether the low-perfusion areas are distributed in a wedge shape along the blood vessels, thereby improving the accuracy of identifying regional low perfusion in subsequent multimodal brain images.

[0076] Furthermore, postoperative tissue hypoperfusion phenotypic information in postoperative multimodal brain imaging was obtained by combining relative cerebral blood flow and maximum delay time. This information included the type and volume of tissue hypoperfusion.

[0077] For the infarcted area, the volume ratio of the area with relative cerebral blood flow less than the flow threshold to the infarcted area is calculated. When the volume ratio is greater than the volume ratio threshold, the tissue hypoperfusion type in the postoperative multimodal brain imaging is determined to be microcirculation no-reflow.

[0078] For example, when there is an area with rCBF < 40% within the infarct area, and the proportion of such area is greater than 10%, it is judged as microcirculation no-reflow. Here, microcirculation no-reflow refers to hypoperfusion at the brain tissue level caused by microvascular failure.

[0079] For non-infarcted areas, when there are hypoperfusion regions within the non-infarcted area, and the region morphology is wedge-shaped and distributed along the blood vessels, the tissue hypoperfusion type in postoperative multimodal brain imaging is determined to be regional hypoperfusion.

[0080] For example, when T exists in the non-infarct region max A hypoperfusion region with a duration of >6 seconds, exhibiting a wedge shape and distribution along the blood vessel, and characterized by an eccentricity >0.7 and a number of connected components <5, is considered a regional hypoperfusion. Regional hypoperfusion is defined as delayed perfusion caused by distal emboli.

[0081] When the tissue hypoperfusion type in postoperative multimodal brain imaging is determined to include both microcirculatory no-reflow and regional hypoperfusion, the tissue hypoperfusion type in postoperative multimodal brain imaging is determined to be a mixed phenotype.

[0082] Furthermore, for regions with low tissue perfusion, the tissue perfusion volume is calculated based on the number of pixels.

[0083] In this implementation, the application specifically determines the relative cerebral blood flow in the infarct area and the maximum delay time in the non-infarct area, thereby matching the perfusion parameters with the low perfusion mechanisms in different regions. By combining the relative cerebral blood flow and the maximum delay time for accurate identification of multimodal brain images, it can effectively distinguish the low perfusion types in the infarct area and the non-infarct area. This solves the problem that the traditional eTICI score cannot distinguish low perfusion of different mechanisms and has large inter-observer variability, providing a standardized, mechanism-specific assessment tool.

[0084] S103. Construct a multidimensional feature vector using clinical feature information, preoperative perfusion feature parameters, postoperative perfusion feature parameters, and postoperative tissue hypoperfusion phenotype information. Input the multidimensional feature vector into the prognostic prediction model to obtain the patient prognostic prediction status information output by the prognostic prediction model.

[0085] The multidimensional feature vector is input into the pre-trained prognostic prediction model to obtain the predicted prognostic status information of the patient output by the prognostic prediction model. Among them, the predicted prognostic status information of the patient is the index information used to characterize the development of tissue hypoperfusion after surgery.

[0086] Optionally, the prediction results of tissue hypoperfusion include predicted functional status information, predicted mortality risk information, and predicted complication status information after a preset time period post-surgery.

[0087] The prognostic prediction model is trained using multidimensional feature vector samples of historical patients and their corresponding prognostic results.

[0088] For example, the prognostic prediction model is a decision tree model. The decision tree model consists of four layers, specifically including a root node, two layers of intermediate decision nodes, and leaf nodes.

[0089] The root node takes a multidimensional feature vector as input and performs a binary branch judgment on the postoperative tissue low perfusion phenotype information in the multidimensional feature vector. It is divided into two parallel branches: high risk of low perfusion delay and low risk of low perfusion delay.

[0090] For branches with a high risk of low perfusion delay, a binary branch determination is performed based on clinical characteristics, preoperative perfusion parameters, and postoperative perfusion parameters, resulting in two predicted tissue low perfusion outcomes, which are then output as leaf nodes. Similarly, for branches with a low risk of low perfusion delay, a binary branch determination is performed based on clinical characteristics, preoperative perfusion parameters, and postoperative perfusion parameters, resulting in two predicted tissue low perfusion outcomes, which are then output as leaf nodes.

[0091] For example, the decision tree model can be a Random Forest model with a tree depth of 5-10 and a learning rate of 0.01-0.1; the decision tree model can be an Extreme Gradient Boosting Tree (XGBoost) model with 100-500 trees; or the decision tree model can be a Lightweight Gradient Boosting Tree (LightGBM) model with 31 leaves and a bagging_fraction of 0.8.

[0092] The steps for training a decision tree model include: The dataset for the pre-decision tree model is constructed by collecting multidimensional feature vector samples from historical patients before or after thromboembolization as feature data, and collecting the postoperative status development results 90 days after the procedure as feature labels. The dataset is constructed by combining the feature data and feature labels.

[0093] Divide the dataset into a training set and a validation set. For example, use 80% of the dataset as the training set and 20% as the validation set.

[0094] With the goal of minimizing the binary cross-entropy loss, we optimize the core hyperparameters of the decision tree model using 5-fold cross-validation. Specifically, the training set is randomly divided into 5 mutually exclusive subsets with equal sample sizes. One subset is used as the validation set, and the remaining 4 subsets are used as the training set. The model is trained, and the validation set performance is calculated. This process is repeated 5 times to obtain 5 sets of validation set performance results. The average value is taken as the final performance of the hyperparameter combination. All hyperparameter combinations to be optimized are traversed, and the hyperparameter combination with the best average performance is selected.

[0095] The decision tree model is trained using gradient descent. Specifically, the weights and biases are updated using gradient descent until the model converges, at which point training stops and is validated using a validation set.

[0096] Finally, the system combines the prognostic predictions and visualizations of patients with tissue hypoperfusion to generate and present a tissue hypoperfusion report. Specifically, the system automatically generates 90-day brain tissue-level hypoperfusion risk, high / low risk stratification, and potential hemorrhagic transformation risk.

[0097] This application uses microvascular no-reflow and regional hypoperfusion phenotypes as core features, combining perfusion parameters and clinical indicators to construct an end-to-end machine learning model. This model provides objective and quantitative probability outputs, avoiding subjective bias. By comprehensively considering patients' clinical characteristics and multimodal imaging information before and after thrombectomy, it addresses the issues of single data sources and heterogeneity, improving prediction accuracy, generalization, and real-time performance. This helps clinicians identify high-risk patients, guide early intervention or adjuvant therapy, ultimately improving patient functional recovery and reducing mortality.

[0098] To achieve the above embodiments, this disclosure also proposes a prognostic device for tissue hypoperfusion after thrombectomy.

[0099] Figure 2 This is a schematic diagram of a prognostic device for post-thrombectomy tissue hypoperfusion, provided in an embodiment of this disclosure. The device can be implemented by software and / or hardware, and is generally integrated into a computing device. Figure 2 As shown, the prognostic device for predicting tissue hypoperfusion after thrombectomy includes: The acquisition module 201 is used to acquire the patient's clinical characteristics, preoperative multimodal brain images before thrombectomy, and postoperative multimodal brain images after thrombectomy.

[0100] Analysis module 202 is used to determine preoperative perfusion characteristic parameters based on preoperative multimodal brain images, and to determine postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on postoperative multimodal brain images.

[0101] The prediction module 203 is used to construct a multidimensional feature vector using clinical feature information, preoperative perfusion feature parameters, postoperative perfusion feature parameters, and postoperative tissue hypoperfusion phenotype information. The multidimensional feature vector is input into the prognostic prediction model to obtain the predicted prognostic status information of the patient output by the prognostic prediction model. The prognostic prediction model is trained using multidimensional feature vector samples of historical patients and corresponding prognostic results. The predicted prognostic status information of the patient is an indicator used to characterize the development of tissue hypoperfusion after surgery.

[0102] In one possible implementation, the acquisition module 201 includes: The data acquisition submodule is used to collect multidimensional feature information of historical patients based on preset multidimensional features, and to collect information on the postoperative status development of historical patients; among which, multidimensional features include a variety of basic clinical features, disease-related features, NIHSS scores, and reperfusion grades.

[0103] The analysis submodule is used to perform correlation analysis on multidimensional feature information and corresponding postoperative status development results information, and calculate the correlation coefficient.

[0104] The filtering submodule is used to filter target features from multidimensional features that meet the correlation criteria, and construct clinical features.

[0105] The first determination submodule is used to collect patients' clinical characteristic information based on clinical features.

[0106] In one possible implementation, the analysis module 202 includes: The segmentation submodule is used to segment preoperative multimodal brain images based on the signal intensity of pixels in the preoperative multimodal brain images, thereby obtaining infarct areas and non-infarct areas.

[0107] The second determination submodule is used to determine the preoperative perfusion characteristic parameters of the whole region based on the change of contrast agent concentration over time in the infarct and non-infarct areas of preoperative multimodal brain imaging. The third determination submodule is used to determine the preoperative perfusion characteristic parameters of the target region based on the change of contrast agent concentration in the infarct area over time in preoperative multimodal brain imaging; among which, the preoperative perfusion characteristic parameters include the mean maximum delay time, the mean cerebral blood flow, and the mean cerebral blood volume.

[0108] In one possible implementation, the analysis module 202 includes: The segmentation submodule is used to segment postoperative multimodal brain images based on the signal intensity of pixels in the images, thereby obtaining infarcted and non-infarcted areas.

[0109] The second determination submodule is used to determine the postoperative perfusion characteristic parameters of the overall region based on the changes in contrast agent concentration over time in the infarct and non-infarct areas of postoperative multimodal brain imaging.

[0110] The third determination submodule is used to determine the postoperative perfusion characteristic parameters of the target region based on the change in contrast agent concentration over time in the infarct area of ​​postoperative multimodal brain imaging. These postoperative perfusion characteristic parameters include the mean maximum delay time, mean cerebral blood flow, and mean cerebral blood volume.

[0111] In one possible implementation, the analysis module 202 includes: The fourth determination submodule is used to determine the relative cerebral blood flow based on the infarct area and the maximum delay time based on the non-infarct area.

[0112] The calculation submodule is used to combine relative cerebral blood flow and maximum delay time to obtain postoperative tissue hypoperfusion phenotype information in postoperative multimodal brain images; wherein, the postoperative tissue hypoperfusion phenotype information includes tissue hypoperfusion type and tissue hypoperfusion volume.

[0113] The prognostic device for post-thrombectomy tissue hypoperfusion provided in this embodiment can execute the prognostic method for post-thrombectomy tissue hypoperfusion provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects of the execution method.

[0114] To implement the above embodiments, this disclosure also proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the prognostic prediction method for post-thrombectomy tissue hypoperfusion in the above embodiments.

[0115] Figure 3 This is a schematic diagram of the structure of a computing device provided in an embodiment of the present disclosure.

[0116] The following is a detailed reference. Figure 3 The diagram illustrates a structural schematic suitable for implementing the computing device 300 in the embodiments of this disclosure. The computing device 300 in the embodiments of this disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The computing device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0117] like Figure 3As shown, the computing device 300 may include a processor (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 302 or a program loaded from memory 308 into random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the computing device 300. The processor 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0118] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows computing device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 A computing device 300 with various devices is shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0119] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from memory 308, or installed from ROM 302. When the computer program is executed by processor 301, it performs the functions defined in the prognostic prediction method for post-thrombectomy tissue hypoperfusion according to embodiments of this disclosure.

[0120] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0121] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0122] The aforementioned computer-readable medium may be included in the aforementioned computing device; or it may exist independently and not assembled into the computing device.

[0123] The aforementioned computer-readable medium carries one or more programs, which, when executed by the computing device, cause the computing device to perform the aforementioned prognostic method for tissue hypoperfusion after thrombectomy.

[0124] The computing device can be programmed with computer program code in one or more programming languages ​​or a combination thereof to perform the operations of this disclosure. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0125] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0126] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units are not, in some cases, intended to limit the specific unit.

[0127] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0128] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0129] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0130] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0131] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative examples of implementing the claims.

Claims

1. A method for predicting the prognosis of tissue hypoperfusion after thrombectomy, characterized in that, The method includes: Collect patients' clinical characteristics, preoperative multimodal brain images before thrombectomy, and postoperative multimodal brain images after thrombectomy; Preoperative perfusion characteristic parameters were determined based on the preoperative multimodal brain images, and postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information were determined based on the postoperative multimodal brain images. A multidimensional feature vector is constructed using the clinical feature information, the preoperative perfusion feature parameters, the postoperative perfusion feature parameters, and the postoperative tissue hypoperfusion phenotype information. The multidimensional feature vector is then input into a prognostic prediction model to obtain the predicted prognostic status information of the patient output by the prognostic prediction model. The prognostic prediction model is trained using multidimensional feature vector samples of historical patients and corresponding prognostic results. The predicted prognostic status information of the patient is an indicator used to characterize the development of tissue hypoperfusion after surgery.

2. The prognostic prediction method for tissue hypoperfusion after thrombectomy according to claim 1, characterized in that, The determination of preoperative perfusion characteristic parameters based on the preoperative multimodal brain imaging includes: According to the preset segmentation strategy, the preoperative multimodal brain images are divided into multiple target regions; Preoperative perfusion characteristic parameters of the whole region were determined based on the change of contrast agent concentration over time in the whole region of preoperative multimodal brain imaging. Preoperative perfusion characteristic parameters of the target region were determined based on the change of contrast agent concentration over time in each target region using preoperative multimodal brain imaging. The preoperative perfusion characteristic parameters include the average time to peak flow, the average cerebral blood flow, and the average cerebral blood volume.

3. The prognostic prediction method for tissue hypoperfusion after thrombectomy according to claim 1, characterized in that, The determination of postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multimodal brain images includes: According to the preset segmentation strategy, the postoperative multimodal brain images are divided into multiple target regions; The postoperative perfusion characteristic parameters of the whole region were determined based on the change of contrast agent concentration over time in the whole region of postoperative multimodal brain imaging. Postoperative perfusion characteristic parameters of the target region were determined based on the change of contrast agent concentration over time in each target region using postoperative multimodal brain imaging. The postoperative perfusion characteristic parameters include the average time to peak flow, the average cerebral blood flow, and the average cerebral blood volume.

4. The prognostic prediction method for tissue hypoperfusion after thrombectomy according to claim 1, characterized in that, The determination of postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multimodal brain images includes: The multimodal brain images are divided based on the signal intensity of pixels in the multimodal brain images to obtain infarct areas and non-infarct areas; The relative cerebral blood flow is determined based on the infarct area, and the peak time is determined based on the non-infarct area; The postoperative tissue hypoperfusion phenotype information in the multimodal brain images is obtained by combining the relative cerebral blood flow and the time to peak flow; wherein, the postoperative tissue hypoperfusion phenotype information includes tissue hypoperfusion type and tissue hypoperfusion volume.

5. The prognostic prediction method for tissue hypoperfusion after thrombectomy according to claim 1, characterized in that, The collection of patients' clinical characteristics information includes: Based on preset multidimensional features, multidimensional feature information of historical patients is collected, and the postoperative status development information of the historical patients is collected; wherein, the multidimensional features include a variety of basic clinical features, disease-related features, NIHSS scores, and reperfusion grades; A correlation analysis was performed on the multidimensional feature information and the corresponding postoperative status development results information, and the correlation coefficient was calculated. Target features that satisfy the correlation condition are selected from the multidimensional features to construct clinical features; The patient's clinical characteristic information is collected based on the aforementioned clinical characteristics.

6. The prognostic prediction method for tissue hypoperfusion after thrombectomy according to claim 1, characterized in that, The preoperative multimodal brain imaging includes CT angiography and CT perfusion imaging, and the postoperative multimodal brain imaging includes CT perfusion imaging and magnetic resonance diffusion-weighted imaging.

7. The prognostic prediction method for tissue hypoperfusion after thrombectomy according to any one of claims 1-6, characterized in that, The patient prognosis prediction status information includes prediction function status information, prediction mortality risk information, and prediction complication status information after a preset time period post-surgery.

8. A prognostic device for tissue hypoperfusion after thrombectomy, characterized in that, The device includes: The acquisition module is used to collect patients' clinical characteristics, preoperative multimodal brain images before thrombectomy, and postoperative multimodal brain images after thrombectomy. The analysis module is used to determine preoperative perfusion characteristic parameters based on the preoperative multimodal brain images, and to determine postoperative perfusion characteristic parameters and postoperative tissue hypoperfusion phenotype information based on the postoperative multimodal brain images. The prediction module is used to construct a multidimensional feature vector using the clinical feature information, the preoperative perfusion feature parameters, the postoperative perfusion feature parameters, and the postoperative tissue hypoperfusion phenotype information. The multidimensional feature vector is then input into the prognostic prediction model to obtain the predicted prognostic status information of the patient output by the prognostic prediction model. The prognostic prediction model is trained using multidimensional feature vector samples of historical patients and corresponding prognostic results. The predicted prognostic status information of the patient is an indicator used to characterize the development of tissue hypoperfusion after surgery.

9. A computing device, characterized in that, The computing device includes: a processor; a memory for storing executable instructions of the processor; the processor for reading the executable instructions from the memory and executing the instructions to implement the prognostic prediction method for tissue hypoperfusion after thrombectomy as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the prognostic method for tissue hypoperfusion after thrombectomy as described in any one of claims 1 to 7.

Images