Clinical prognosis prediction method, device, equipment and storage medium before thrombectomy
Through the infarction zone, penumbra volume and location information based on image data, combined with clinical prognostic model, the insufficient screening accuracy and radiation of thrombectomy surgery in the existing technology are solved, and the high level of patients with acute ischemic stroke is achieved Accuracy filtering.
Patent Information
- Application Number
- CN202411108140.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-08-13
AI Technical Summary
The existing screening methods for thrombectomy surgery in patients with acute ischemic stroke are insufficient, and CTP and DSC technologies can produce radiation and are not suitable for some patients, especially for patients with renal insufficiency, contrast allergies or pediatric patients.
Using a method based on three-dimensional T1 weighted image, apparent diffusion coefficient image, cerebral blood flow image and arterial passage time image, combined with key area images, the volume of infarction zone, gray matter semidark zone and white matter semidark zone is determined, and feature vectors are generated. And input a pre-established clinical prognosis prediction model to predict the clinical prognosis results of thrombectomy surgery.
It improves the prediction accuracy of clinical prognostic results of thrombectomy surgery, and can accurately screen patients with high benefits-risk ratio without radiation and contrast agent, and is suitable for a variety of patient groups.
Smart Images

Figure CN118645210B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent diagnosis of acute ischemic stroke, and in particular to a method, device, equipment and storage medium for predicting clinical prognosis before thrombectomy. Background Art
[0002] Acute ischemic stroke is a common cerebrovascular disease that has always been a serious threat to human health. Its morbidity, disability, and mortality rates are very high. Thrombectomy for patients with acute ischemic stroke can save the ischemic penumbra and prevent it from further developing into infarction. However, thrombectomy also has the risk of bleeding, so it is particularly important to screen out patients who are likely to benefit from thrombectomy.
[0003] Currently, patients with a high benefit-risk ratio are usually selected for thrombectomy by combining information such as core infarct volume, ratio of low perfusion area to core infarct, ischemic penumbra volume, etc. Among them, the low perfusion area is usually obtained from computed tomography perfusion (CTP) images or magnetic resonance dynamic susceptibility contrast (DSC) images.
[0004] However, the method of combining information such as core infarct volume, ratio of low perfusion area to core infarct, ischemic penumbra volume, etc. to determine patients with a high benefit-risk ratio has great limitations, and its accuracy needs to be improved. In addition, CTP and DSC produce radiation and require injection of contrast agents, which will have a certain impact on patients. They are not suitable for patients with renal insufficiency, patients allergic to contrast agents, and children. Summary of the invention
[0005] In view of the above-mentioned problems in the prior art, the object of the present invention is to provide a method, device, equipment and storage medium for predicting clinical prognosis before thrombectomy surgery, which can improve the prediction accuracy of clinical prognosis results of thrombectomy surgery.
[0006] In order to solve the above problems, the present invention provides a method for predicting clinical prognosis before thrombectomy surgery, comprising:
[0007] Step S1, obtaining a three-dimensional T1 weighted image, an apparent diffusion coefficient image, a cerebral blood flow image, and an arterial transit time image of the target patient's brain;
[0008] Step S2, determining the infarct volume, the key gray matter penumbra volume, and the key white matter penumbra volume of the target patient's brain based on the three-dimensional T1 weighted image, the apparent diffusion coefficient image, the cerebral blood flow image, and the arterial transit time image in combination with the key area image;
[0009] Step S3, generating a feature vector of the target patient's brain based on the infarct area volume, the key gray matter penumbra volume, and the key white matter penumbra volume;
[0010] Step S4, inputting the feature vector into a pre-established clinical prognosis prediction model to obtain the clinical prognosis result of the target patient's brain thrombectomy surgery.
[0011] Furthermore, the step S1 comprises:
[0012] Acquiring diffusion-weighted image data and arterial spin labeling perfusion image data of the brain of a target patient;
[0013] determining the apparent diffusion coefficient image based on the diffusion weighted image data;
[0014] The cerebral blood flow image and the arterial transit time image are determined based on the arterial spin labeling perfusion image data.
[0015] Furthermore, the step S2 comprises:
[0016] Step S201, determining the infarct area of the brain of the target patient based on the apparent diffusion coefficient image;
[0017] Step S202, registering the infarct area to the standard brain space to obtain the infarct area in the standard brain space;
[0018] Step S203, calculating the volume of the infarct area in the standard brain space.
[0019] Furthermore, the step S2 further includes:
[0020] Step S204, generating a gray-white matter segmentation image of the target patient's brain based on the three-dimensional T1 weighted image;
[0021] Step S205, determining a gray matter hypoperfusion area and a white matter hypoperfusion area according to the gray-white matter segmentation image and the cerebral blood flow image, and registering the gray matter hypoperfusion area and the white matter hypoperfusion area to a standard brain space, respectively, to obtain the gray matter hypoperfusion area and the white matter hypoperfusion area of the standard brain space;
[0022] Step S206, determining a first key gray matter penumbra based on the key area image, the gray matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space;
[0023] Step S207, determining a first key white matter penumbra based on the key area image, the white matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space;
[0024] Step S208, determining the gray matter long artery passage time zone and the white matter long artery passage time zone according to the gray-white matter segmentation image and the artery passage time image, and registering the gray matter long artery passage time zone and the white matter long artery passage time zone to the standard brain space, respectively, to obtain the gray matter long artery passage time zone and the white matter long artery passage time zone of the standard brain space;
[0025] Step S209, determining a second key gray matter penumbra based on the key area image, the gray matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space;
[0026] Step S210, determining a second key white matter penumbra based on the key area image, the white matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space;
[0027] Step S211, respectively calculating the volumes of the first focused gray matter penumbra, the second focused gray matter penumbra, the first focused white matter penumbra, and the second focused white matter penumbra.
[0028] Furthermore, the step S205 includes:
[0029] Registering the cerebral blood flow image to a three-dimensional T1 weighted image space to obtain a cerebral blood flow image in the three-dimensional T1 weighted image space;
[0030] Determine the cerebral blood flow value corresponding to each voxel in the gray-white matter segmentation image according to the cerebral blood flow image in the three-dimensional T1 weighted image space;
[0031] Acquire voxels whose corresponding cerebral blood flow values are less than a first threshold value in the gray matter region of the gray-white matter segmentation image to generate the gray matter hypoperfusion region;
[0032] Acquire voxels whose corresponding cerebral blood flow values are less than a second threshold value in the white matter region of the gray-white matter segmentation image to generate the white matter hypoperfusion region;
[0033] The step S208 includes:
[0034] Registering the artery transit time image to the three-dimensional T1 weighted image space to obtain the artery transit time image in the three-dimensional T1 weighted image space;
[0035] Determine the arterial transit time value corresponding to each voxel in the gray-white matter segmentation image according to the arterial transit time image in the three-dimensional T1 weighted image space;
[0036] Acquire voxels whose corresponding arterial transit time values are greater than a third threshold value in the gray matter region of the gray-white matter segmentation image, and generate the gray matter long artery transit time region;
[0037] The voxels whose corresponding artery transit time values are greater than a fourth threshold value in the white matter region of the gray-white matter segmentation image are acquired to generate the white matter long artery transit time region.
[0038] Furthermore, the step S206 includes:
[0039] After taking a difference set between the gray matter hypoperfusion area of the standard brain space and the infarct area of the standard brain space, taking an intersection set with the key area in the key area image to obtain the first key gray matter penumbra;
[0040] The step S207 includes:
[0041] After taking a difference set between the white matter hypoperfusion area in the standard brain space and the infarct area in the standard brain space, taking an intersection set with the key area in the key area image to obtain the first key white matter penumbra;
[0042] The step S209 includes:
[0043] After taking a difference set between the gray matter long artery passing time zone of the standard brain space and the infarct area of the standard brain space, taking an intersection set with the key area in the key area image to obtain the second key gray matter penumbra;
[0044] The step S210 includes:
[0045] The second key white matter penumbra is obtained by taking a difference set between the white matter long artery passing time zone of the standard brain space and the infarct area of the standard brain space, and then taking an intersection with the key area in the key area image.
[0046] Furthermore, the step S3 comprises:
[0047] Acquire clinical information of a target patient, wherein the clinical information includes basic information and diagnosis and treatment information of the target patient;
[0048] The feature vector is generated according to the clinical information, the infarct volume, the focused gray matter penumbra volume and the focused white matter penumbra volume.
[0049] Furthermore, the method further comprises pre-establishing the clinical prognosis prediction model, and the process of establishing the clinical prognosis prediction model comprises:
[0050] Acquire a training sample data set, wherein the training sample data set includes three-dimensional T1 weighted image samples, apparent diffusion coefficient image samples, cerebral blood flow image samples, and arterial transit time image samples of the brains of multiple patients who have successfully undergone thrombectomy surgery, as well as corresponding postoperative improvement conditions;
[0051] Determine the sample infarct volume, the sample key gray matter penumbra volume, and the sample key white matter penumbra volume of the corresponding patient's brain based on the three-dimensional T1 weighted image samples, the apparent diffusion coefficient image samples, the cerebral blood flow image samples, and the arterial transit time image samples of each patient's brain in combination with the key area image;
[0052] Generate a feature vector sample corresponding to the patient's brain based on the sample infarct area volume, the sample key gray matter penumbra volume, and the sample key white matter penumbra volume;
[0053] The clinical prognosis prediction model is established by taking the characteristic vector samples of each patient's brain as the independent variable and the corresponding postoperative improvement as the dependent variable.
[0054] Furthermore, the method for generating the key area image includes:
[0055] Acquire a patient data set, wherein the patient data set includes multiple apparent diffusion coefficient images of the brains of patients and corresponding scoring data;
[0056] Determining the infarct area of the brain of the corresponding patient based on each of the apparent diffusion coefficient images, and registering the infarct area to the standard brain space to obtain the infarct area of the standard brain space;
[0057] Based on the infarct area of the standard brain space of each patient's brain, the score data is used as the dependent variable, and whether the voxel is in the infarct area is used as the independent variable, voxel-based injury-symptom mapping is performed to obtain the area significantly correlated with the score data;
[0058] The key area image is determined according to the area significantly related to the scoring data.
[0059] Another aspect of the present invention provides a clinical prognosis prediction device before thrombectomy surgery, comprising:
[0060] An image acquisition module, used to acquire a three-dimensional T1-weighted image, an apparent diffusion coefficient image, a cerebral blood flow image, and an arterial transit time image of the target patient's brain;
[0061] a volume determination module, configured to determine the infarct volume, the gray matter penumbra volume, and the white matter penumbra volume of the target patient's brain based on the three-dimensional T1 weighted image, the apparent diffusion coefficient image, the cerebral blood flow image, and the arterial transit time image in combination with a key area image;
[0062] A feature vector generating module, used for generating a feature vector of the target patient's brain based on the infarct area volume, the key gray matter penumbra volume and the key white matter penumbra volume;
[0063] The prediction module is used to input the feature vector into a pre-established clinical prognosis prediction model to obtain the clinical prognosis result of the thrombectomy surgery on the target patient's brain.
[0064] On the other hand, the present invention provides an electronic device, including a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the clinical prognosis prediction method before thrombectomy surgery as described above.
[0065] On the other hand, the present invention provides a computer-readable storage medium, which stores at least one instruction or at least one program. The at least one instruction or the at least one program is loaded and executed by a processor to implement the clinical prognosis prediction method before thrombectomy surgery as described above.
[0066] Due to the above technical solution, the present invention has the following beneficial effects:
[0067] According to the clinical prognosis prediction method before thrombectomy surgery in an embodiment of the present invention, the infarct area is determined based on diffusion weighted imaging (DWI) image data, the key gray matter penumbra and the key white matter penumbra are determined based on arterial spin labeling (ASL) perfusion image data, and a pre-established clinical prognosis prediction model is used to predict the clinical prognosis of thrombectomy surgery in the target patient's brain based on the infarct area volume, the key gray matter penumbra volume, and the key white matter penumbra volume. This method introduces the position information and tissue type information of the penumbra on the basis of the core infarct area volume and the penumbra volume, which can improve the prediction accuracy of the clinical prognosis of thrombectomy surgery, thereby accurately screening out stroke patients with a high benefit-risk ratio without radiation or contrast agents. BRIEF DESCRIPTION OF THE DRAWINGS
[0068] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments or prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.
[0069] Figure 1 It is a schematic diagram of an implementation environment provided by an embodiment of the present invention;
[0070] Figure 2 is a flow chart of a method for predicting clinical prognosis before thrombectomy surgery provided by one embodiment of the present invention;
[0071] Figure 3is a flow chart of a method for predicting clinical prognosis before thrombectomy provided by another embodiment of the present invention;
[0072] Figure 4 is a schematic diagram of a key gray matter penumbra provided by an embodiment of the present invention;
[0073] Figure 5 is a schematic diagram of a key area image provided by an embodiment of the present invention;
[0074] Figure 6 is a flow chart of a method for establishing a clinical prognosis prediction model provided by one embodiment of the present invention;
[0075] Figure 7 It is a schematic structural diagram of a clinical prognosis prediction device before thrombectomy surgery provided by one embodiment of the present invention;
[0076] Figure 8 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0077] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0078] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or equipment that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.
[0079] Reference Manual Attached Figure 1 , which shows a schematic diagram of an implementation environment provided by an embodiment of the present invention. Figure 1As shown, the implementation environment may include at least one medical scanning device 110 and a computer device 120. The computer device 120 and each medical scanning device 110 may be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment of the present invention.
[0080] Among them, the medical scanning device 110 can be but is not limited to a magnetic resonance imaging (MRI) device, etc., and the computer device 120 can be but is not limited to various servers, personal computers, laptops, smart phones, tablet computers and portable wearable devices. The server can be an independent server or a server cluster or distributed system composed of multiple servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs) and big data and artificial intelligence platforms.
[0081] In the embodiment of the present invention, the medical scanning device 110 can collect images of the target patient's brain based on MRI technology to obtain corresponding three-dimensional T1 weighted images, DWI image data and ASL perfusion image data. The computer device 120 can obtain the three-dimensional T1 weighted images, DWI image data and ASL perfusion image data scanned by the medical scanning device 110, and determine the clinical prognosis of the thrombectomy surgery in the target patient's brain through the clinical prognosis prediction method before thrombectomy surgery provided by the embodiment of the present invention, so as to provide doctors with reference to guide the screening of patients with high benefit-risk ratio for thrombectomy surgery.
[0082] It should be noted that Figure 1 It is just an example. Those skilled in the art will understand that although Figure 1 Only one medical scanning device 110 is shown, but this does not constitute a limitation on the embodiment of the present invention, and more or less medical scanning devices 110 than shown in the figure may be included.
[0083] Reference Manual Attached Figure 2 , which shows the process of a clinical prognosis prediction method before thrombectomy surgery provided by an embodiment of the present invention, which can be applied to Figure 1 In the computer device 120, specifically Figure 2 As shown, the method may include the following steps:
[0084] Step S1, obtaining a three-dimensional T1 weighted image, an apparent diffusion coefficient image, a cerebral blood flow image, and an arterial transit time image of the target patient's brain.
[0085] In an embodiment of the present invention, MRI technology can be used to collect images of the target patient's brain to obtain a corresponding three-dimensional T1 weighted image, which refers to an image of the T1 longitudinal relaxation of the protruding tissue. The target patient may be a patient suffering from acute ischemic stroke.
[0086] In the embodiment of the present invention, the magnetic resonance DWI imaging technology can be used to collect the image of the target patient's brain to obtain the corresponding DWI image data, and then the DWI image data can be used to calculate the apparent diffusion coefficient (ADC) image.
[0087] In an embodiment of the present invention, the ASL perfusion imaging technology can be used to collect images of the target patient's brain to obtain corresponding ASL perfusion image data, and then the ASL perfusion image data is used to calculate the cerebral blood flow (CBF) image and the arterial transit time (ATT) image. The ASL perfusion image data may be multi-labeled post-delay (Multi-PLD) ASL perfusion image data.
[0088] Specifically, the step S1 may include: acquiring diffusion-weighted image data and arterial spin labeling perfusion image data of the target patient's brain; determining the apparent diffusion coefficient image based on the diffusion-weighted image data; and determining the cerebral blood flow image and the arterial transit time image based on the arterial spin labeling perfusion image data.
[0089] It should be noted that the ADC image may be determined based on the DWI image data using an existing method, or the CBF image and the ATT image may be determined based on the ASL perfusion image data using an existing method, which will not be described in detail in the embodiments of the present invention.
[0090] It can be understood that compared with CTP and DSC, ASL perfusion has the advantages of no radiation and no need for contrast agents. That is to say, the data acquisition process of the embodiment of the present invention is radiation-free and contrast-free, and can be applied to patients with renal insufficiency, patients allergic to contrast agents, and children.
[0091] Step S2, based on the three-dimensional T1 weighted image, the apparent diffusion coefficient image, the cerebral blood flow image and the arterial transit time image, combined with the key area image, determine the infarct area volume, the key gray matter penumbra volume and the key white matter penumbra volume of the target patient's brain.
[0092] In practical applications, as the research is carried out more extensively and deeply, researchers have found that the clinical outcomes of patients with acute ischemic stroke are not only related to the infarct volume, but also to the location of the infarct. Specifically, various areas of the brain are related to different functions. When areas involving some basic functions are damaged, more serious symptoms will result; while infarctions occurring in other areas have relatively little impact on basic living abilities. Therefore, in addition to the infarct volume, the location of the infarction is also related to the severity of the stroke and the long-term functional outcomes. In view of this, in an embodiment of the present invention, the location information of the penumbra can be introduced when predicting the patient's clinical prognosis to select patients for surgery.
[0093] In addition, since different tissue types have different tolerances to ischemia, for example, gray matter is more susceptible to ischemia than white matter. At the same time, since normal gray matter and white matter have differences in perfusion parameters, it is more necessary to distinguish them. Therefore, in the embodiment of the present invention, the tissue type of the penumbra can be introduced into the clinical prognosis prediction model to improve the accuracy of the prediction results.
[0094] In an embodiment of the present invention, the infarct volume, as well as the focused gray matter penumbra volume and the focused white matter penumbra volume that can reflect the penumbra location information and tissue type can be determined based on the three-dimensional T1 weighted image, the ADC image, the CBF image, and the ATT image in combination with the focused area image. The focused gray matter penumbra volume may include but is not limited to the focused gray matter penumbra volume related to CBF and the focused gray matter penumbra volume related to ATT, and the focused white matter penumbra volume may include but is not limited to the focused white matter penumbra volume related to CBF and the focused white matter penumbra volume related to ATT.
[0095] In a possible embodiment, in conjunction with the reference specification, Figure 3 , the step S2 may include:
[0096] Step S201: determining the infarct area in the brain of the target patient based on the apparent diffusion coefficient image.
[0097] Specifically, the infarct area of the target patient's brain can be obtained according to the ADC image of the target patient's brain, or by combining the DWI and ADC images using existing methods, which will not be described in detail in the embodiments of the present invention.
[0098] In a possible embodiment, the infarction area of the target patient's brain can be determined by combining the ADC image and the CBF image, or by combining the ADC image, the CBF image, and the cerebral blood volume image (CBV) and the ATT image calculated based on the ASL perfusion image data. The specific determination method can refer to the prior art, and the embodiments of the present invention will not be repeated here.
[0099] Step S202: registering the infarct area to the standard brain space to obtain the infarct area in the standard brain space.
[0100] Specifically, the infarct region may be registered to the standard brain space using an existing registration method to obtain the infarct region of the standard brain space, and the embodiments of the present invention will not be described in detail herein.
[0101] Step S203, calculating the volume of the infarct area in the standard brain space.
[0102] Specifically, after obtaining the infarct area of the standard brain space, the number of voxels in the infarct area can be determined, and then the volume of the infarct area in the standard brain space can be obtained by multiplying the volume of each voxel.
[0103] In a possible embodiment, in conjunction with the reference specification, Figure 3 , the step S2 may further include:
[0104] Step S204: generating a gray-white matter segmentation image of the target patient's brain based on the three-dimensional T1 weighted image.
[0105] Specifically, an existing segmentation method, such as a segmentation method based on a deep neural network, a segmentation method based on a classifier, etc., can be used to segment the three-dimensional T1 weighted image to obtain a gray-white matter segmentation image, which will not be described in detail in the embodiments of the present invention. The gray-white matter segmentation image may include a gray matter area and a white matter area.
[0106] Step S205, determining the gray matter hypoperfusion area and the white matter hypoperfusion area according to the gray-white matter segmentation image and the cerebral blood flow image, and registering the gray matter hypoperfusion area and the white matter hypoperfusion area to the standard brain space, respectively, to obtain the gray matter hypoperfusion area and the white matter hypoperfusion area of the standard brain space.
[0107] Specifically, the hypoperfusion area of the target patient's brain can be determined based on the CBF image, and the hypoperfusion area located in the gray matter area, i.e., the gray matter hypoperfusion area, and the hypoperfusion area located in the white matter area, i.e., the white matter hypoperfusion area, can be determined based on the gray-white matter segmentation image.
[0108] Optionally, step S205 may include: registering the cerebral blood flow image to the three-dimensional T1 weighted image space to obtain the cerebral blood flow image in the three-dimensional T1 weighted image space; determining the cerebral blood flow value corresponding to each voxel in the gray-white matter segmentation image according to the cerebral blood flow image in the three-dimensional T1 weighted image space; acquiring the voxels in the gray matter area of the gray-white matter segmentation image whose corresponding cerebral blood flow value is less than a first threshold, and generating the gray matter hypoperfusion area; acquiring the voxels in the white matter area of the gray-white matter segmentation image whose corresponding cerebral blood flow value is less than a second threshold, and generating the white matter hypoperfusion area.
[0109] Specifically, in order to simplify the subsequent processing process, the CBF image may be registered to the three-dimensional T1 weighted image space using an existing registration method to obtain a CBF image in the three-dimensional T1 weighted image space.
[0110] Specifically, the gray-white matter hypoperfusion area can be determined according to the gray-white matter area and the CBF value corresponding to each voxel. The gray-matter hypoperfusion area can include voxels in the gray matter area whose corresponding CBF value is less than a first threshold, and the white-matter hypoperfusion area can include voxels in the white matter area whose corresponding CBF value is less than a second threshold. Wherein, both the first threshold and the second threshold can be preset and adjusted according to clinical experience. For example, the first threshold can be set to 30 mL / (100 g·min), that is, the voxels in the gray matter area with a CBF value less than 30 mL / (100 g·min) form a gray-matter hypoperfusion area, and the second threshold can be set to 15 mL / (100 g·min), that is, the voxels in the white matter area with a CBF value less than 15 mL / (100 g·min) form a white-matter hypoperfusion area.
[0111] In a possible embodiment, a relative threshold may also be used, that is, the first threshold and the second threshold are determined according to the CBF value of the target patient's brain.
[0112] Specifically, when the low perfusion area is located only on one side of the target subject's brain, a first preset ratio of the normal side cerebral gray matter CBF mean value can be used as the first threshold, and a second preset ratio of the normal side cerebral white matter CBF mean value can be used as the second threshold. The first preset ratio and the second preset ratio can be preset and adjusted according to clinical experience. For example, 40% of the normal side cerebral gray matter CBF mean value can be used as the first threshold, and 40% of the normal side cerebral white matter CBF mean value can be used as the second threshold. The embodiment of the present invention does not impose specific restrictions on this.
[0113] Specifically, when there are low perfusion areas on both sides of the target subject's brain, the third preset ratio of the whole brain gray matter CBF mean can be used as the first threshold, and the fourth preset ratio of the whole brain white matter CBF mean can be used as the second threshold. The third preset ratio and the fourth preset ratio can be preset and adjusted according to clinical experience. For example, 40% of the whole brain gray matter CBF mean can be used as the first threshold, and 40% of the whole brain white matter CBF mean can be used as the second threshold. The embodiment of the present invention does not impose specific restrictions on this.
[0114] In practical applications, the midline of the target object's brain can be determined first, and the target object's brain can be divided into two sides according to the midline. The difference between the CBF values on both sides of the target object's brain can be compared using the T test method. If there is no significant difference between the CBF values on both sides of the target object's brain, it can be determined that there are low perfusion areas on both sides of the target object's brain. If there is a significant difference between the CBF values on both sides of the target object's brain, it can be determined that the low perfusion area is only located on one side of the target object's brain, and the side with significantly higher CBF values can be determined as the normal side.
[0115] It can be understood that by adopting a relative threshold method to determine the low perfusion area judgment threshold according to the CBF value of the target patient's brain, more targeted threshold information can be obtained, thereby obtaining a more accurate gray and white matter low perfusion area.
[0116] Specifically, after determining the gray matter hypoperfusion area and the white matter hypoperfusion area, existing registration methods can be used to respectively register them to the standard brain space to obtain the gray matter hypoperfusion area and the white matter hypoperfusion area of the standard brain space, so as to eliminate the influence of the differences in brain volume of different patients on the clinical prognosis results.
[0117] Step S206: determining a first key gray matter penumbra based on the key region image, the gray matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space.
[0118] Specifically, the step S206 may include: taking a difference set between the gray matter hypoperfusion area of the standard brain space and the infarct area of the standard brain space, and then taking an intersection set with the key area in the key area image to obtain the first key gray matter penumbra. The first key gray matter penumbra is a CBF-related key gray matter penumbra, and the key area image is a key area image of the standard brain space.
[0119] For example, Figure 4 As shown, the gray matter low perfusion area 41 and the infarction area 42 can be subtracted and then intersected with the key area 43 to obtain the CBF-related key gray matter penumbra 44.
[0120] Step S207: determining a first key white matter penumbra based on the key region image, the white matter hypoperfusion region of the standard brain space, and the infarct region of the standard brain space.
[0121] Specifically, step S207 may include: taking a difference set between the white matter hypoperfusion area of the standard brain space and the infarct area of the standard brain space, and then taking an intersection set with the key area in the key area image to obtain the first key white matter penumbra, wherein the first key white matter penumbra is a CBF-related key white matter penumbra.
[0122] Step S208, determining the gray matter long artery passage time zone and the white matter long artery passage time zone according to the gray-white matter segmentation image and the artery passage time image, and aligning the gray matter long artery passage time zone and the white matter long artery passage time zone to the standard brain space, respectively, to obtain the gray matter long artery passage time zone and the white matter long artery passage time zone of the standard brain space.
[0123] Specifically, the long ATT area of the target patient's brain can be determined based on the ATT image, and the long ATT area located in the gray matter area, i.e., the gray matter long ATT area, and the long ATT area located in the white matter area, i.e., the white matter long ATT area, can be determined respectively based on the gray and white matter segmentation image.
[0124] Optionally, step S208 may include: registering the artery transit time image to the three-dimensional T1 weighted image space to obtain the artery transit time image in the three-dimensional T1 weighted image space; determining the artery transit time value corresponding to each voxel in the gray-white matter segmentation image according to the artery transit time image in the three-dimensional T1 weighted image space; acquiring the voxels whose corresponding artery transit time values in the gray matter area of the gray-white matter segmentation image are greater than a third preset threshold, and generating the gray matter long artery transit time zone; acquiring the voxels whose corresponding artery transit time values in the white matter area of the gray-white matter segmentation image are greater than a fourth preset threshold, and generating the white matter long artery transit time zone.
[0125] Specifically, in order to simplify the subsequent processing process, the ATT image can be registered to the three-dimensional T1 weighted image space using an existing registration method to obtain the ATT image in the three-dimensional T1 weighted image space.
[0126] Specifically, the gray matter long ATT area can be determined according to the gray matter area and the ATT value corresponding to each voxel. The gray matter long ATT area may include voxels in the gray matter area whose corresponding ATT value is greater than the third threshold, and the white matter long ATT area may include voxels in the white matter area whose corresponding ATT value is greater than the fourth threshold. Wherein, the third threshold and the fourth threshold can be pre-set and adjusted according to clinical experience. For example, the third threshold can be set to 1.8s, that is, the voxels in the gray matter area with an ATT value greater than 1.8s form a gray matter long ATT area, and the fourth threshold can be set to 2s, that is, the voxels in the white matter area with an ATT value greater than 2s form a white matter long ATT area.
[0127] In a possible embodiment, a relative threshold value may also be used, that is, the third threshold value and the fourth threshold value are determined according to the ATT value of the target patient's brain. The specific determination method can refer to the method for determining the first threshold value and the second threshold value when the relative threshold value is used in step S205, and the embodiment of the present invention will not be repeated here.
[0128] It can be understood that by adopting a relative threshold method to determine the long ATT area judgment threshold according to the ATT value of the target patient's brain, more targeted threshold information can be obtained, thereby obtaining a more accurate gray and white matter long ATT area.
[0129] Specifically, after determining the gray matter long ATT area and the white matter long ATT area, they can be respectively aligned to the standard brain space using existing alignment methods to obtain the gray matter long ATT area and the white matter long ATT area of the standard brain space, so as to eliminate the impact of the differences in brain volume of different patients on the clinical prognosis results.
[0130] Step S209, determining a second key gray matter penumbra based on the key area image, the gray matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space.
[0131] Specifically, the step S209 may include: taking a difference set between the gray matter long artery passing time zone of the standard brain space and the infarct area of the standard brain space, and then taking an intersection set with the key area in the key area image to obtain the second key gray matter penumbra. The second key gray matter penumbra is an ATT-related key gray matter penumbra.
[0132] Step S210: determining a second key white matter penumbra based on the key area image, the white matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space.
[0133] Specifically, the step S210 may include: taking a difference set between the white matter long artery passing time zone of the standard brain space and the infarct zone of the standard brain space, and then taking an intersection set with the key area in the key area image to obtain the second key white matter penumbra, wherein the second key white matter penumbra is an ATT-related key white matter penumbra.
[0134] Step S211, respectively calculating the volumes of the first focused gray matter penumbra, the second focused gray matter penumbra, the first focused white matter penumbra, and the second focused white matter penumbra.
[0135] Specifically, after obtaining the first focused gray matter penumbra, the second focused gray matter penumbra, the first focused white matter penumbra and the second focused white matter penumbra, the number of voxels of the first focused gray matter penumbra, the second focused gray matter penumbra, the first focused white matter penumbra and the second focused white matter penumbra can be determined respectively, and then multiplied by the volume of each voxel to obtain their respective volumes.
[0136] In the embodiment of the present invention, the key area image can be pre-generated based on a patient data set including multiple patients' brain apparent diffusion coefficient images and corresponding scoring data. Figure 5 The key area image may be an image of a brain area including a standard brain space and a key area in the brain area (such as area 51 in the figure). The key area refers to an area where, once infarction occurs, more serious consequences will occur, such as a decline in the ability to take care of oneself.
[0137] It should be noted that the formation of the infarction zone is a process, first ischemia forms the ischemic zone, and then slowly develops into infarction. If the thrombectomy is successful, it can prevent the ischemic zone from further developing into the infarction zone. Therefore, if ischemia occurs in the above-mentioned key areas, it is believed that thrombectomy may prevent it from becoming an infarction zone, and the reperfusion of blood flow can slowly restore the corresponding function of this area, which may improve the symptoms.
[0138] In a possible embodiment, the method for generating the key area image may include the following steps: acquiring a patient data set, the patient data set including multiple apparent diffusion coefficient images of the patient's brain and corresponding scoring data; determining the infarct area of the corresponding patient's brain based on each of the apparent diffusion coefficient images, and aligning the infarct area to the standard brain space to obtain the infarct area of the standard brain space; based on the infarct area of the standard brain space of each of the patient's brains, using the scoring data as the dependent variable and whether the voxel is in the infarct area as the independent variable, performing voxel-based injury-symptom mapping to obtain an area significantly correlated with the scoring data; and determining the key area image based on the area significantly correlated with the scoring data.
[0139] The scoring data may include, but are not limited to, the National Institutes of Health Stroke Scale (NIHSS) score and the 3-6 month Modified Rankin Scale (mRS) score.
[0140] Specifically, multiple acute ischemic stroke patient data may be collected to form a patient data set, and the patient data may include ADC images of the patient's brain, NIHSS scores, and 3-6 month mRS scores. The ADC images in the patient data set are artifact-free images.
[0141] Exemplarily, DWI image data, NIHSS scores, and 3-6 month mRS scores of the brains of at least 300 patients with acute ischemic stroke within 24-72 hours can be collected, and the DWI image data with excellent scanning quality, no motion artifacts, and calculated ADC images without artifacts and their corresponding NIHSS scores and mRS scores can be selected. And for each patient, an ADC image is calculated based on the DWI image data. The ADC images of the brains of the selected multiple patients and their corresponding NIHSS scores and mRS scores are used as patient data for generating the key area image.
[0142] Specifically, for each patient, the existing method can be used to obtain the infarction area of the corresponding patient's brain based on the ADC image of the patient's brain, or by combining DWI and ADC images, and then the existing registration method can be used to register it to the standard brain space to obtain the infarction area of the standard brain space. The embodiment of the present invention will not repeat the process of determining the infarction area and the registration process. So far, each patient's data includes the infarction area of the standard brain space, NIHSS score and mRS score.
[0143] Specifically, the NIHSS score and the mRS score can be used as dependent variables, and whether the voxel is in the infarction area can be used as an independent variable to perform voxel-based lesion-symptom mapping (VLSM). The specific steps are: the voxels at a certain spatial position in the brain of all patients are divided into two groups: infarction and non-infarction, and the differences in NIHSS scores between the groups are compared by methods including but not limited to T tests to obtain t values and p values. After repeating the above operations for voxels at all spatial positions, multiple comparison corrections are performed to obtain areas significantly correlated with NIHSS. The mRS score is subjected to the same operation as the NIHSS score to obtain areas significantly correlated with mRS. Finally, the area significantly correlated with NIHSS and the area significantly correlated with mRS can be taken as the union of the key area to obtain an image of the key area.
[0144] It should be noted that the specific process of calculating the t value and the p value and the specific process of performing the multiple comparison correction can refer to the existing technical content, and the embodiments of the present invention will not be repeated here.
[0145] In a possible embodiment, the scoring data may also include only the NIHSS score. In this case, the area significantly related to the NIHSS score may be directly used as the key area to obtain a key area image.
[0146] It can be understood that by pre-generating key area images based on ADC images of the brains of multiple patients and corresponding scoring data, the infarction location area related to the severity of stroke and long-term functional outcome can be accurately determined.
[0147] Step S3, generating a feature vector of the target patient's brain based on the infarct area volume, the key gray matter penumbra volume, and the key white matter penumbra volume.
[0148] In an embodiment of the present invention, the feature vector of the target patient's brain can be directly composed of the infarct volume, the key gray matter penumbra volume, and the key white matter penumbra volume. Exemplarily, the key gray matter penumbra volume can include the CBF-related key gray matter penumbra volume and the ATT-related key gray matter penumbra volume, and the key white matter penumbra volume can include the CBF-related key white matter penumbra volume and the ATT-related key white matter penumbra volume. Then, the infarct volume, the CBF-related key gray matter penumbra volume, the CBF-related key white matter penumbra volume, the ATT-related key gray matter penumbra volume, and the ATT-related key white matter penumbra volume can be directly used to form the feature vector.
[0149] In a possible embodiment, the characteristic vector of the target patient's brain may also include clinical information of the target patient related to the improvement after thrombectomy, and the clinical information may include but is not limited to basic information and diagnosis and treatment information of the target patient. The basic information may include but is not limited to information such as age, and the diagnosis and treatment information may include but is not limited to information such as whether the target patient has diabetes and information such as the target patient's blood pressure.
[0150] Specifically, the step S3 may include: acquiring clinical information of the target patient; and generating the feature vector according to the clinical information, the infarct area volume, the key gray matter penumbra volume, and the key white matter penumbra volume.
[0151] Specifically, basic information such as the age of the target patient can be obtained from the diagnosis and treatment system or from various types of acquired image data, and diagnosis and treatment information such as whether the target patient suffers from diabetes and blood pressure can be obtained from the diagnosis and treatment system.
[0152] Specifically, the feature vector of the target patient's brain may be composed of the clinical information, the infarct volume, the focused gray matter penumbra volume, and the focused white matter penumbra volume. Exemplarily, the clinical information may include age, whether suffering from diabetes, and blood pressure, the focused gray matter penumbra volume may include a focused gray matter penumbra volume related to CBF and a focused gray matter penumbra volume related to ATT, and the focused white matter penumbra volume may include a focused white matter penumbra volume related to CBF and a focused white matter penumbra volume related to ATT. Then, the feature vector may be composed of age, whether suffering from diabetes, blood pressure, infarct volume, focused gray matter penumbra volume related to CBF, focused white matter penumbra volume related to CBF, focused gray matter penumbra volume related to ATT, and focused white matter penumbra volume related to ATT.
[0153] It can be understood that by combining the clinical information of the target patient, the infarct area information of the target patient's brain, the location information of the penumbra and the tissue type information, etc., to predict the clinical prognosis of thrombectomy in the target patient's brain, the accuracy of the clinical prognosis of thrombectomy can be improved, thereby accurately screening out stroke patients with a high benefit-risk ratio.
[0154] It should be noted that the above-mentioned embodiment of predicting the clinical prognosis of thrombectomy surgery using clinical information, infarct area volume, key gray matter penumbra volume and key white matter penumbra volume is only an example. The embodiment of the present invention does not specifically limit the type of characteristic data used to predict the clinical prognosis of thrombectomy surgery. In actual applications, those skilled in the art can determine it according to the actual situation.
[0155] In some possible embodiments, only the calculated infarct volume, the focused gray matter penumbra volume, and the focused white matter penumbra volume may be used as feature data. In other words, the feature vector of the target patient's brain may only include the infarct volume, the focused gray matter penumbra volume, and the focused white matter penumbra volume.
[0156] Step S4, inputting the feature vector into a pre-established clinical prognosis prediction model to obtain the clinical prognosis result of the target patient's brain thrombectomy surgery.
[0157] In an embodiment of the present invention, the clinical prognosis of thrombectomy can be predicted based on the characteristic vector of the target patient's brain and using a pre-established clinical prognosis prediction model to obtain the clinical prognosis of thrombectomy in the target patient's brain, thereby helping to screen patients who can benefit more from the surgery. Among them, the clinical prognosis of thrombectomy can include significant improvement and no significant improvement. If the clinical prognosis of thrombectomy is significant improvement, it means that if thrombectomy is performed, the patient's symptoms will be significantly improved after the thrombectomy, so the target patient can be subjected to thrombectomy; if the clinical prognosis of thrombectomy is no significant improvement, it means that if thrombectomy is performed, the patient's symptoms will not be significantly improved after the thrombectomy, so there is no need to perform thrombectomy on the target patient.
[0158] In an embodiment of the present invention, the clinical prognosis prediction model may include but is not limited to a logistic regression model, and the embodiment of the present invention does not impose specific restrictions on the structure of the model. Exemplarily, when the clinical prognosis prediction model is a logistic regression model, the clinical prognosis result of the thrombectomy surgery may be a value between [0,1]. When the clinical prognosis result of the thrombectomy surgery is less than or equal to the fifth threshold, it can be considered that the clinical prognosis result of the thrombectomy surgery is not significantly improved. When the clinical prognosis result of the thrombectomy surgery is greater than the fifth threshold, it can be considered that the clinical prognosis result of the thrombectomy surgery is significantly improved. Among them, the fifth threshold is a value between 0 and 1. The fifth threshold can be pre-set according to actual conditions, for example, it can be set to 0.5. The embodiment of the present invention does not impose specific restrictions on this.
[0159] In an embodiment of the present invention, a training sample data set including three-dimensional T1 weighted image samples, ADC image samples, CBF image samples and ATT image samples of the brains of multiple patients who have successfully undergone thrombectomy surgery, as well as corresponding postoperative improvement conditions can be obtained in advance, and the sample infarct area volume, sample key gray matter penumbra volume and sample key white matter penumbra volume of each patient's brain can be determined in combination with the key area image to construct the corresponding feature vector sample, and the clinical prognosis prediction model is pre-established with the feature vector sample as the independent variable and the corresponding postoperative improvement condition as the dependent variable. The method for establishing the clinical prognosis prediction model will be described in detail later.
[0160] It should be noted that the type of feature vectors in the feature vector sample used to establish the clinical prognosis prediction model can be the same as the type of feature vectors in the feature vector constructed when predicting the clinical prognosis results of thrombectomy surgery.
[0161] In summary, according to the clinical prognosis prediction method before thrombectomy surgery in an embodiment of the present invention, the infarct area is determined based on DWI image data, the key gray matter penumbra and the key white matter penumbra are determined based on ASL perfusion image data, and a pre-established clinical prognosis prediction model is used to predict the clinical prognosis of thrombectomy surgery in the target patient's brain based on the infarct area volume, the key gray matter penumbra volume, and the key white matter penumbra volume. This method introduces the position information and tissue type information of the penumbra on the basis of the core infarct area volume and the penumbra volume, which can improve the prediction accuracy of the clinical prognosis of thrombectomy surgery, thereby accurately screening out stroke patients with a high benefit-risk ratio without radiation or contrast agents.
[0162] Reference Manual Attached Figure 6 , which shows the process of establishing a clinical prognosis prediction model provided by an embodiment of the present invention, specifically as follows Figure 6 As shown, the method may include the following steps:
[0163] S610, obtaining a training sample data set, wherein the training sample data set includes three-dimensional T1 weighted image samples, apparent diffusion coefficient image samples, cerebral blood flow image samples and arterial transit time image samples of the brains of multiple patients who have successfully undergone thrombectomy surgery, as well as corresponding postoperative improvement conditions.
[0164] In an embodiment of the present invention, data of multiple patients with acute ischemic stroke who have successfully undergone thrombectomy can be collected to form a training sample data set, and the patient data may include three-dimensional T1 weighted image samples, ADC image samples, CBF image samples, ATT image samples, and postoperative improvement of the patient's brain before surgery. Among them, the three-dimensional T1 weighted image samples, ADC image samples, CBF image samples, and ATT image samples in the training sample data set are all artifact-free images. The postoperative improvement can be determined based on the NIHSS scores of the corresponding patient's brain before and after surgery. Specifically, a decrease of 4 points or more in the NIHSS score before and after surgery or a NIHSS score of zero after surgery is considered to be a significant improvement after surgery, otherwise it is considered that there is no significant improvement after surgery.
[0165] Exemplarily, three-dimensional T1 weighted image samples, DWI image data samples, ASL perfusion image data samples, and NIHSS scores before and after the operation can be collected from at least 300 patients with acute ischemic stroke who have successfully undergone thrombectomy, and the patient data of the DWI image data samples with excellent scanning quality and no motion artifacts, and the calculated ADC image samples without artifacts, and the calculated CBF image samples and ATT image samples without artifacts are selected. And for each patient, the ADC image samples are calculated according to the DWI image data samples, the CBF image samples and ATT image samples are calculated according to the ASL perfusion image data samples, and the corresponding postoperative improvement is determined according to the NIHSS scores before and after the operation. The three-dimensional T1 weighted image samples, ADC image samples, CBF image samples, ATT image samples of the selected multiple patients' brains and the corresponding postoperative improvement are used as patient data for establishing a clinical prognosis prediction model.
[0166] The specific implementation method of determining the ADC image sample based on the DWI image data sample, and determining the CBF image sample and the ATT image sample based on the ASL perfusion image data sample can refer to the prior art, and the embodiments of the present invention will not be repeated here. The ASL perfusion image data sample can be a Multi-PLD ASL perfusion image data sample.
[0167] In a possible embodiment, the patient data for establishing a clinical prognosis prediction model may also include clinical information of the patient, and the clinical information may include but is not limited to basic information and diagnosis and treatment information of the patient. The basic information may include but is not limited to information such as age, and the diagnosis and treatment information may include but is not limited to information such as whether the patient has diabetes and information such as the patient's blood pressure.
[0168] S620, determining a sample infarct area volume, a sample key gray matter penumbra volume, and a sample key white matter penumbra volume of the corresponding patient's brain based on the three-dimensional T1 weighted image samples, the apparent diffusion coefficient image samples, the cerebral blood flow image samples, and the arterial transit time image samples of each patient's brain in combination with the key area image.
[0169] In the embodiment of the present invention, the specific content of step S620 can be referred to Figure 2 In the illustrated embodiment, in step S2, the infarct volume, the key gray matter penumbra volume, and the key white matter penumbra volume of the target patient's brain are determined based on the three-dimensional T1 weighted image, ADC image, CBF image, and ATT image in combination with the key area image. The specific contents of the embodiment of the present invention will not be repeated here.
[0170] S630: Generate a feature vector sample corresponding to the patient's brain based on the sample infarct area volume, the sample key gray matter penumbra volume, and the sample key white matter penumbra volume.
[0171] In the embodiment of the present invention, the specific content of step S630 can be referred to Figure 2 The specific contents of generating the characteristic vector of the target patient's brain based on the infarct area volume, the key gray matter penumbra volume and the key white matter penumbra volume in step S3 in the illustrated embodiment will not be described in detail in the embodiment of the present invention.
[0172] S640, establishing the clinical prognosis prediction model using the characteristic vector samples of each patient's brain as an independent variable and the corresponding postoperative improvement as a dependent variable.
[0173] In the embodiment of the present invention, a clinical prognosis prediction model can be constructed with the feature vector sample as an independent variable and the corresponding postoperative improvement as a dependent variable. The clinical prognosis prediction model can include but is not limited to a logistic regression model, and the embodiment of the present invention does not specifically limit the structure of the model.
[0174] For example, a logistic regression model can be constructed with age, diabetes, blood pressure, infarct volume, CBF-related gray matter penumbra volume, CBF-related white matter penumbra volume, ATT-related gray matter penumbra volume, and ATT-related white matter penumbra volume as independent variables, and with significant improvement after surgery as the dependent variable. Specifically, the logistic regression model is shown in the following formula:
[0175] Among them, x is the eigenvector, ω is the parameter, b is the bias, and y indicates whether there is significant improvement after surgery.
[0176] Specifically, according to each feature vector sample and its corresponding postoperative improvement, the parameters can be adjusted to obtain the model parameters that minimize the loss function. And according to the prediction performance of the model, such as Youden index, F1 value, etc., the thresholds of the infarct area and the low perfusion area can be adjusted, and finally a clinical prognosis prediction model with good prediction performance is obtained. Among them, the loss function and the prediction performance type of the model can be pre-set according to actual needs, and the embodiment of the present invention does not impose specific restrictions on this.
[0177] It should be noted that, in the embodiment of the present invention, other specific contents of the model building process can be referred to Figures 2 to 5 The specific content of the method provided in the illustrated embodiment may refer to the prior art, and the embodiments of the present invention will not be described in detail here.
[0178] It can be understood that by using the volume of the infarct area obtained based on DWI image data, the volume of the key gray matter penumbra obtained based on ASL perfusion image data, and the volume of the key white matter penumbra obtained based on the ASL perfusion image data as independent variables, a clinical prognosis prediction model is established. On the basis of the core infarct area volume and the penumbra volume, the position information and tissue type information of the penumbra are introduced, which can improve the predictive effect of the clinical prognosis prediction model, thereby improving the predictive accuracy of the clinical prognosis results of thrombectomy surgery, and then accurately screening out stroke patients with a high benefit-risk ratio without radiation and contrast agents.
[0179] Reference Manual Attached Figure 7 , which shows the structure of a clinical prognosis prediction device 700 before thrombectomy surgery provided by an embodiment of the present invention. Figure 7 As shown, the device 700 may include:
[0180] An image acquisition module 710 is used to acquire a three-dimensional T1-weighted image, an apparent diffusion coefficient image, a cerebral blood flow image, and an arterial transit time image of the target patient's brain;
[0181] A volume determination module 720 is used to determine the infarct volume, the gray matter penumbra volume and the white matter penumbra volume of the target patient's brain based on the three-dimensional T1 weighted image, the apparent diffusion coefficient image, the cerebral blood flow image and the arterial transit time image in combination with the key area image;
[0182] A feature vector generating module 730, configured to generate a feature vector of the target patient's brain based on the infarct volume, the focused gray matter penumbra volume, and the focused white matter penumbra volume;
[0183] The prediction module 740 is used to input the feature vector into a pre-established clinical prognosis prediction model to obtain the clinical prognosis result of the thrombectomy surgery on the target patient's brain.
[0184] In a possible embodiment, the device 700 may further include a model building module for pre-building the clinical prognosis prediction model; the model building module may include:
[0185] A sample data acquisition unit is used to acquire a training sample data set, wherein the training sample data set includes a plurality of three-dimensional T1 weighted image samples, apparent diffusion coefficient image samples, cerebral blood flow image samples and arterial transit time image samples of the brain of patients who have successfully undergone thrombectomy surgery, as well as corresponding postoperative improvement conditions;
[0186] a sample volume determination unit, configured to determine a sample infarct area volume, a sample key gray matter penumbra volume, and a sample key white matter penumbra volume of the corresponding patient's brain based on the three-dimensional T1 weighted image samples, the apparent diffusion coefficient image samples, the cerebral blood flow image samples, and the arterial transit time image samples of each patient's brain, respectively, in combination with a key area image;
[0187] A sample feature generating unit, used for generating a feature vector sample corresponding to the patient's brain based on the sample infarct area volume, the sample key gray matter penumbra volume and the sample key white matter penumbra volume;
[0188] The model building unit is used to build the clinical prognosis prediction model by taking the characteristic vector samples of each patient's brain as the independent variable and the corresponding postoperative improvement as the dependent variable.
[0189] In a possible embodiment, the apparatus 700 may further include a key area image generation module, which is used to pre-generate the key area image; the key area image generation module may include:
[0190] A data set acquisition unit, used to acquire a patient data set, wherein the patient data set includes multiple apparent diffusion coefficient images of the brains of patients and corresponding scoring data;
[0191] an infarct area determination unit, configured to determine the infarct area of the corresponding patient's brain based on each of the apparent diffusion coefficient images, and register the infarct area to a standard brain space to obtain the infarct area of the standard brain space;
[0192] A significant region determination unit is used to perform voxel-based injury-symptom mapping based on the infarct region of the standard brain space of each patient's brain, taking the score data as a dependent variable and whether the voxel is in the infarct region as an independent variable, to obtain a region significantly correlated with the score data;
[0193] A key area image determining unit is used to determine the key area image according to the area significantly related to the scoring data.
[0194] It should be noted that the device provided in the above embodiment, when implementing its functions, only uses the division of the above functional modules as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiment and the corresponding method embodiment belong to the same concept, and the specific implementation process is detailed in the corresponding method embodiment, which will not be repeated here.
[0195] An embodiment of the present invention also provides an electronic device, which includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the clinical prognosis prediction method before thrombectomy surgery provided in the above-mentioned method embodiment.
[0196] The memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.
[0197] In a specific embodiment, Figure 8 The hardware structure diagram of an electronic device for implementing the clinical prognosis prediction method before thrombectomy provided by an embodiment of the present invention is shown. The electronic device may be a computer terminal, a mobile terminal or other device. The electronic device may also participate in or include the clinical prognosis prediction device before thrombectomy provided by an embodiment of the present invention. Figure 8 As shown, the electronic device 800 may include a memory 810 of one or more computer-readable storage media, a processor 820 of one or more processing cores, an input unit 830, a display unit 840, a radio frequency (RF) circuit 850, a wireless fidelity (WiFi) module 860, and a power supply 870. Those skilled in the art will appreciate that Figure 8 The electronic device structure shown in the figure does not constitute a limitation on the electronic device 800, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.
[0198] The memory 810 can be used to store software programs and modules. The processor 820 executes various functional applications and data processing by running or executing the software programs and modules stored in the memory 810 and calling the data stored in the memory 810. The memory 810 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the electronic device, etc. In addition, the memory 810 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 810 can also include a memory controller to provide the processor 820 with access to the memory 810.
[0199] The processor 820 is the control center of the electronic device 800. It uses various interfaces and lines to connect various parts of the entire electronic device. By running or executing software programs and / or modules stored in the memory 810, and calling data stored in the memory 810, it executes various functions of the electronic device 800 and processes data, thereby monitoring the electronic device 800 as a whole. The processor 820 can be a central processing unit, or other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application-specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.
[0200] The input unit 830 may be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control. Specifically, the input unit 830 may include a touch-sensitive surface 831 and other input devices 832. Specifically, the touch-sensitive surface 831 may include but is not limited to a touch pad or a touch screen, and the other input devices 832 may include but are not limited to one or more of a physical keyboard, a function key (such as a volume control key, a switch key, etc.), a trackball, a mouse, a joystick, etc.
[0201] The display unit 840 may be used to display information input by a user or information provided to a user and various graphical user interfaces of the electronic device, which may be composed of graphics, text, icons, videos, and any combination thereof. The display unit 840 may include a display panel 841, which may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), or the like.
[0202] The RF circuit 850 can be used for receiving and sending signals during information transmission or calls. In particular, after receiving the downlink information of the base station, it is handed over to one or more processors 820 for processing; in addition, the data related to the uplink is sent to the base station. Generally, the RF circuit 850 includes but is not limited to an antenna, at least one amplifier, a tuner, one or more oscillators, a user identity module (SIM) card, a transceiver, a coupler, a low noise amplifier (Low Noise Amplifier, LNA), a duplexer, etc. In addition, the RF circuit 850 can also communicate with the network and other devices through wireless communication. The wireless communication can use any communication standard or protocol, including but not limited to Global System of Mobilecommunication (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), email, Short Messaging Service (SMS), etc.
[0203] WiFi is a short-range wireless transmission technology. The electronic device 800 can help users send and receive emails, browse web pages, and access streaming media through the WiFi module 860. It provides users with wireless broadband Internet access. Figure 8 A WiFi module 860 is shown, but it is understandable that it is not an essential component of the electronic device 800 and can be omitted as needed without changing the essence of the invention.
[0204] The electronic device 800 also includes a power supply 870 (such as a battery) for supplying power to each component. Preferably, the power supply can be logically connected to the processor 820 through a power management system, so that the power management system can manage charging, discharging, and power consumption management. The power supply 870 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components.
[0205] It should be noted that, although not shown, the electronic device 800 may also include a Bluetooth module, etc., which will not be described in detail here.
[0206] An embodiment of the present invention also provides a computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to implementing a clinical prognosis prediction method before thrombectomy surgery. The at least one instruction or the at least one program is loaded and executed by the processor to implement the clinical prognosis prediction method before thrombectomy surgery provided by the above-mentioned method embodiment.
[0207] Optionally, in an embodiment of the present invention, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.
[0208] One embodiment of the present invention further provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the clinical prognosis prediction method before thrombectomy surgery provided in the above-mentioned various optional implementation examples.
[0209] It should be noted that the sequence of the embodiments of the present invention described above is for description only and does not represent the advantages and disadvantages of the embodiments. The above describes specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0210] 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 device 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.
[0211] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.
[0212] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for predicting clinical prognosis before thrombectomy surgery, characterized in that: include: Step S1, obtaining a three-dimensional T1 weighted image, an apparent diffusion coefficient image, a cerebral blood flow image, and an arterial transit time image of the target patient's brain; Step S2, determining the infarct volume, the key gray matter penumbra volume, and the key white matter penumbra volume of the target patient's brain based on the three-dimensional T1 weighted image, the apparent diffusion coefficient image, the cerebral blood flow image, and the arterial transit time image in combination with the key area image; Step S3, generating a feature vector of the target patient's brain based on the infarct area volume, the key gray matter penumbra volume, and the key white matter penumbra volume; Step S4, inputting the feature vector into a pre-established clinical prognosis prediction model to obtain a clinical prognosis result of the target patient's brain thrombectomy surgery; Wherein, the step S2 comprises: Step S201, determining the infarct area of the brain of the target patient based on the apparent diffusion coefficient image; Step S202, registering the infarct area to the standard brain space to obtain the infarct area in the standard brain space; Step S203, calculating the volume of the infarct area in the standard brain space; Step S204, generating a gray-white matter segmentation image of the target patient's brain based on the three-dimensional T1 weighted image; Step S205, determining a gray matter hypoperfusion area and a white matter hypoperfusion area according to the gray-white matter segmentation image and the cerebral blood flow image, and registering the gray matter hypoperfusion area and the white matter hypoperfusion area to a standard brain space, respectively, to obtain the gray matter hypoperfusion area and the white matter hypoperfusion area of the standard brain space; Step S206, determining a first key gray matter penumbra based on the key area image, the gray matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space; Step S207, determining a first key white matter penumbra based on the key area image, the white matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space; Step S208, determining the gray matter long artery passage time zone and the white matter long artery passage time zone according to the gray-white matter segmentation image and the artery passage time image, and registering the gray matter long artery passage time zone and the white matter long artery passage time zone to the standard brain space, respectively, to obtain the gray matter long artery passage time zone and the white matter long artery passage time zone of the standard brain space; Step S209, determining a second key gray matter penumbra based on the key area image, the gray matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space; Step S210, determining a second key white matter penumbra based on the key area image, the white matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space; Step S211, respectively calculating the volumes of the first focused gray matter penumbra, the second focused gray matter penumbra, the first focused white matter penumbra, and the second focused white matter penumbra.
2. The method according to claim 1, characterized in that: The step S1 comprises: Acquiring diffusion-weighted image data and arterial spin labeling perfusion image data of the brain of a target patient; determining the apparent diffusion coefficient image based on the diffusion weighted image data; The cerebral blood flow image and the arterial transit time image are determined based on the arterial spin labeling perfusion image data.
3. The method according to claim 1, characterized in that The step S205 includes: Registering the cerebral blood flow image to a three-dimensional T1 weighted image space to obtain a cerebral blood flow image in the three-dimensional T1 weighted image space; Determine the cerebral blood flow value corresponding to each voxel in the gray-white matter segmentation image according to the cerebral blood flow image in the three-dimensional T1 weighted image space; Acquire voxels whose corresponding cerebral blood flow values are less than a first threshold value in the gray matter region of the gray-white matter segmentation image to generate the gray matter hypoperfusion region; Acquire voxels whose corresponding cerebral blood flow values are less than a second threshold value in the white matter region of the gray-white matter segmentation image to generate the white matter hypoperfusion region; The step S208 includes: Registering the artery transit time image to the three-dimensional T1 weighted image space to obtain the artery transit time image in the three-dimensional T1 weighted image space; Determine the arterial transit time value corresponding to each voxel in the gray-white matter segmentation image according to the arterial transit time image in the three-dimensional T1 weighted image space; Acquire voxels whose corresponding arterial transit time values are greater than a third threshold value in the gray matter region of the gray-white matter segmentation image, and generate the gray matter long artery transit time region; The voxels whose corresponding artery transit time values are greater than a fourth threshold value in the white matter region of the gray-white matter segmentation image are acquired to generate the white matter long artery transit time region.
4. The method according to claim 1, characterized in that: The step S206 includes: After taking a difference set between the gray matter hypoperfusion area of the standard brain space and the infarct area of the standard brain space, taking an intersection set with the key area in the key area image to obtain the first key gray matter penumbra; The step S207 includes: After taking a difference set between the white matter hypoperfusion area in the standard brain space and the infarct area in the standard brain space, taking an intersection set with the key area in the key area image to obtain the first key white matter penumbra; The step S209 includes: After taking a difference set between the gray matter long artery passing time zone of the standard brain space and the infarct area of the standard brain space, taking an intersection set with the key area in the key area image to obtain the second key gray matter penumbra; The step S210 includes: The second key white matter penumbra is obtained by taking a difference set between the white matter long artery passing time zone of the standard brain space and the infarct area of the standard brain space, and then taking an intersection with the key area in the key area image.
5. The method according to claim 1, characterized in that The step S3 comprises: Acquire clinical information of a target patient, wherein the clinical information includes basic information and diagnosis and treatment information of the target patient; The feature vector is generated according to the clinical information, the infarct volume, the focused gray matter penumbra volume and the focused white matter penumbra volume.
6. The method according to claim 1, characterized in that The method further includes pre-establishing the clinical prognosis prediction model, and the process of establishing the clinical prognosis prediction model includes: Acquire a training sample data set, wherein the training sample data set includes three-dimensional T1 weighted image samples, apparent diffusion coefficient image samples, cerebral blood flow image samples, and arterial transit time image samples of the brains of multiple patients who have successfully undergone thrombectomy surgery, as well as corresponding postoperative improvement conditions; Determine the sample infarct volume, the sample key gray matter penumbra volume, and the sample key white matter penumbra volume of the corresponding patient's brain based on the three-dimensional T1 weighted image samples, the apparent diffusion coefficient image samples, the cerebral blood flow image samples, and the arterial transit time image samples of each patient's brain in combination with the key area image; Generate a feature vector sample corresponding to the patient's brain based on the sample infarct area volume, the sample key gray matter penumbra volume, and the sample key white matter penumbra volume; The clinical prognosis prediction model is established by taking the characteristic vector samples of each patient's brain as the independent variable and the corresponding postoperative improvement as the dependent variable.
7. The method according to claim 1 or 6, characterized in that: The method for generating the key area image includes: Acquire a patient data set, wherein the patient data set includes multiple apparent diffusion coefficient images of the brains of patients and corresponding scoring data; Determining the infarct area of the brain of the corresponding patient based on each of the apparent diffusion coefficient images, and registering the infarct area to the standard brain space to obtain the infarct area of the standard brain space; Based on the infarct area of the standard brain space of each patient's brain, the score data is used as the dependent variable, and whether the voxel is in the infarct area is used as the independent variable, voxel-based injury-symptom mapping is performed to obtain the area significantly correlated with the score data; The key area image is determined according to the area significantly related to the scoring data.
8. A clinical prognosis prediction device before thrombectomy surgery, characterized in that: include: An image acquisition module, used to acquire a three-dimensional T1-weighted image, an apparent diffusion coefficient image, a cerebral blood flow image, and an arterial transit time image of the target patient's brain; A volume determination module is used to determine the infarct volume, the key gray matter penumbra volume and the key white matter penumbra volume of the target patient's brain based on the three-dimensional T1 weighted image, the apparent diffusion coefficient image, the cerebral blood flow image and the arterial transit time image in combination with the key area image, including: Step S201, determining the infarct area of the brain of the target patient based on the apparent diffusion coefficient image; Step S202, registering the infarct area to the standard brain space to obtain the infarct area in the standard brain space; Step S203, calculating the volume of the infarct area in the standard brain space; Step S204, generating a gray-white matter segmentation image of the target patient's brain based on the three-dimensional T1 weighted image; Step S205, determining a gray matter hypoperfusion area and a white matter hypoperfusion area according to the gray-white matter segmentation image and the cerebral blood flow image, and registering the gray matter hypoperfusion area and the white matter hypoperfusion area to a standard brain space, respectively, to obtain the gray matter hypoperfusion area and the white matter hypoperfusion area of the standard brain space; Step S206, determining a first key gray matter penumbra based on the key area image, the gray matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space; Step S207, determining a first key white matter penumbra based on the key area image, the white matter hypoperfusion area of the standard brain space, and the infarct area of the standard brain space; Step S208, determining the gray matter long artery passage time zone and the white matter long artery passage time zone according to the gray-white matter segmentation image and the artery passage time image, and registering the gray matter long artery passage time zone and the white matter long artery passage time zone to the standard brain space, respectively, to obtain the gray matter long artery passage time zone and the white matter long artery passage time zone of the standard brain space; Step S209, determining a second key gray matter penumbra based on the key area image, the gray matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space; Step S210, determining a second key white matter penumbra based on the key area image, the white matter long artery passage time zone of the standard brain space, and the infarct area of the standard brain space; Step S211, respectively calculating the volumes of the first focused gray matter penumbra, the second focused gray matter penumbra, the first focused white matter penumbra, and the second focused white matter penumbra; A feature vector generating module, used for generating a feature vector of the target patient's brain based on the infarct area volume, the key gray matter penumbra volume and the key white matter penumbra volume; The prediction module is used to input the feature vector into a pre-established clinical prognosis prediction model to obtain the clinical prognosis result of the thrombectomy surgery on the target patient's brain.
9. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the clinical prognosis prediction method before thrombectomy surgery as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one instruction or at least one program, and the at least one instruction or at least one program is loaded and executed by the processor to implement the clinical prognosis prediction and detection method before thrombectomy surgery as described in any one of claims 1-7.
Citation Information
Patent Citations
Arterial spin labeling-based cerebral hypoperfusion region accurate quantification method
CN112288705A
Image region segmentation method and device, equipment and storage medium
CN113628207A
Ischemic region segmentation method, device and equipment and storage medium
CN114820602A
Cerebral hemorrhage prognosis prediction method, device, equipment, medium and program product
CN117789992A
Post-thrombectomy effect prediction system
CN117796829A
