Methods and apparatus for analyzing lesion locations in brain diseases

Abstract

The present invention relates to a method and apparatus for analyzing the location of a lesion in a brain disease. A method according to one embodiment of the present invention is a method performed in an electronic device and comprises: a step of aligning a plurality of first medical images so that the brain shape is aligned with a plurality of first medical images of brain regions of patients with vascular occlusion stroke; a step of setting a common region including common voxels that are voxels of lesion locations overlapping in at least N images (where N is a natural number greater than or equal to 2) among the plurality of first medical images; a step of performing clustering on the data of the common region to divide the common region into a plurality of regions, thereby deriving a clustered image in which each divided region is displayed; and a step of analyzing whether the stroke is caused by a small vessel blockage or a large vessel inlet blockage by dividing each region of the stroke area to reflect the pathological cause of a small subcortical cerebral infarction.

Description

Technology Field

[0001] The present invention relates to a technology for analyzing the location of lesions in brain diseases, and more specifically, to a technology for analyzing the detailed location of lesions in strokes. Background Technology

[0003] In cases of brain diseases such as stroke, the "golden time" for treatment is critical. To this end, a rapid and accurate analysis of the patient's condition must be performed based on medical imaging of the brain. In this context, medical imaging refers to images of the patient's internal brain structure, including structural details of the brain region, internal tissues, and fluid flow.

[0004] A technology for analyzing brain diseases based on such medical images is disclosed in KR 10-2012-50379 A (hereinafter referred to as the "prior art"). This prior art relates to a technology for determining whether a subject to diagnosis has a brain disease.

[0005] On the other hand, in the case of cerebral infarctions such as small subcortical infarctions, the treatment method varies depending on the location of the occlusion causing the stroke (i.e., the location of the lesion). For example, the treatment method may differ depending on whether the stroke is caused by an occlusion of a small blood vessel located in a specific part of the brain or by an occlusion at the entrance of a large blood vessel corresponding to the superior vessel from which the small blood vessels differentiate.

[0006] However, conventional technology has limitations, such as merely determining whether a brain disease has occurred or being unable to differentiate between small vessel blockage and blockage at the entrance of large vessels.

[0007] However, the above description merely provides background information regarding the present invention and does not constitute previously disclosed technology. Prior art literature

[0009] (Patent Document 0001) KR 10-2012-50379 A The problem to be solved

[0010] In order to solve the problems of the conventional technology described above, the present invention aims to provide a technology for analyzing the detailed lesion location of a stroke.

[0011] In particular, the present invention aims to provide a technique for analyzing the detailed lesion location for small subcortical cerebral infarctions, such as small subcortical cerebral infarctions, or large vessel branch occlusion cerebral infarctions.

[0012] However, the problems that the present invention aims to solve are not limited to those mentioned above, and other unmentioned problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0014] A method according to an embodiment of the present invention for solving the above-mentioned problem is a method performed in an electronic device, comprising: a step of aligning a plurality of first medical images so that the brain shape is aligned with a plurality of first medical images of brain regions of patients with vascular occlusion stroke; a step of setting a common region including common voxels that are voxels of lesion locations overlapping in at least N images (wherein N is a natural number greater than or equal to 2) among the plurality of first medical images; and a step of performing clustering on the data of the common region to divide the common region into a plurality of regions, thereby deriving a clustered image in which each divided region is displayed.

[0015] A method according to one embodiment of the present invention further includes a step of reducing the dimensionality of the data in the common region between the setting step and the deriving step, and the deriving step may include a step of performing clustering on the data in the common region with reduced data dimensionality.

[0016] A method according to one embodiment of the present invention may further include the step of displaying a template for the clustered image and a second medical image of a brain region of a patient with vascular occlusion stroke who is the subject of treatment, respectively.

[0017] The above-mentioned display step may include a step of displaying such that each clustered region of the template and the stroke region of the second medical image overlap.

[0018] The above-mentioned first medical image may be a medical image of patients who have developed a vascular occlusion stroke due to a blockage of either a first blockage corresponding to a small vessel blockage or a second blockage corresponding to a large vessel entrance blockage.

[0019] A method according to one embodiment of the present invention may further include the step of determining that a patient in the second medical image has developed a vascular occlusion stroke due to a first blockage or a second blockage based on the similarity between each clustered region of the clustered image and the stroke site of the second medical image.

[0020] The above-mentioned judgment step may include a step of determining that the patient in the second medical image has developed a vascular occlusion stroke due to the first blockage when the similarity is above a certain level, and determining that the patient in the second medical image has developed a vascular occlusion stroke due to the second blockage when the similarity is below a certain level.

[0021] The above first medical image may include a 1-1 medical image of patients who developed a vascular occlusive stroke due to a blockage of either a first blockage corresponding to a small vessel blockage or a second blockage corresponding to a large vessel entrance blockage, and a 1-2 medical image of patients who developed a vascular occlusive stroke due to a blockage of the other of the first and second blockages.

[0022] The plurality of first medical images used in the above setting step are the first-1 medical images, and the method according to one embodiment of the present invention may further include a step of determining whether the clustered image is suitable by comparing the clustered image and the first-2 medical images after the above derivation step.

[0023] The above-mentioned judgment step may include a step of determining the suitability of the clustered image based on whether the difference between the stroke area of ​​the first-second medical image and each clustered region of the clustered image is greater than a certain level.

[0024] The above-mentioned setting step includes the step of setting a first common area which is a common area for a plurality of first-1 medical images and the step of setting a second common area which is a common area for a plurality of first-2 medical images, and the above-mentioned deriving step may include the step of deriving a first clustered image which is a clustered image for the data of the first common area and deriving a second clustered image which is a clustered image for the data of the second common area.

[0025] An apparatus according to one embodiment of the present invention includes: a memory storing a plurality of first medical images of brain regions of patients with vascular occlusion stroke; and a control unit that controls the derivation of clustered images to assist in the analysis of a second medical image of a patient with vascular occlusion stroke to be treated, using the plurality of first medical images.

[0026] The control unit can control the registration of the plurality of first medical images so that the brain shape of each first medical image is registered, control the setting of a common region including common voxels that are voxels of overlapping lesion locations in at least N images (wherein N is a natural number greater than or equal to 2) among the plurality of first medical images, and control the derivation of the clustered image in which each divided region is displayed by performing clustering on the data of the common region and dividing the common region into multiple regions.

[0027] The above control unit controls the reduction of the dimensionality of the data in the common area and can control the clustering to be performed on the data in the common area with reduced data dimensionality.

[0028] The above control unit can control the display of a template for clustered images and a second medical image of a brain region of a patient with vascular occlusion stroke, which is the target of treatment, respectively.

[0029] The first medical image above is a medical image of patients who have developed a vascular occlusive stroke due to a blockage of either a first blockage corresponding to a small vessel blockage or a second blockage corresponding to a large vessel entrance blockage, and the control unit can control the determination that the patient in the second medical image has developed a vascular occlusive stroke due to the first blockage or the second blockage based on the similarity between each clustered region of the clustered image and the stroke area of ​​the second medical image.

[0030] The control unit can control the determination that the patient in the second medical image has suffered a vascular occlusion stroke due to the first blockage when the similarity is above a certain level, and that the patient in the second medical image has suffered a vascular occlusion stroke due to the second blockage when the similarity is below a certain level.

[0031] The above first medical image may each include a 1-1 medical image of patients who developed a vascular occlusive stroke due to a blockage of either a first blockage corresponding to a small vessel blockage or a second blockage corresponding to a large vessel blockage, and a 1-2 medical image of patients who developed a vascular occlusive stroke due to a blockage of the other of the first and second blockages.

[0032] The above control unit can control the setting of the common area using a plurality of 1-1 medical images and control the determination of whether the clustered image is suitable by comparing the clustered image and the 1-2 medical image.

[0033] The above control unit can control whether the clustered image is suitable based on whether the difference between the stroke area of ​​the first-second medical image and each clustered region of the clustered image is greater than a certain level.

[0034] The control unit controls the setting of a first common area, which is a common area for a plurality of first-1 medical images, and a second common area, which is a common area for a plurality of first-2 medical images; controls the derivation of a first clustered image, which is a clustered image for the data of the first common area, and controls the derivation of a second clustered image, which is a clustered image for the data of the second common area; and controls the first and second clustered images and a second medical image, which is a brain region of a patient with vascular occlusion stroke who is the subject of treatment, to be displayed on a display, respectively.

[0035] An apparatus according to another embodiment of the present invention is an apparatus for assisting in the analysis of a second medical image of a brain region of a patient with vascular occlusion stroke to be treated, using a plurality of first medical images of a brain region of a patient with vascular occlusion stroke derived from a plurality of first medical images of a patient with vascular occlusion stroke, and comprises: a memory storing the clustered image and the second medical image; and a control unit controlling the display of a template for the clustered image and the second medical image on a display, respectively.

[0036] The clustered image may be an image prepared through a process of matching the plurality of first medical images so that the brain shape of each first medical image is matched, a process of setting a common region including common voxels that are voxels of overlapping lesion locations in at least N images (wherein N is a natural number greater than or equal to 2) among the plurality of first medical images, and a process of performing clustering on the data of the common region to divide the common region into multiple regions and thereby deriving the clustered image in which each divided region is displayed. Effects of the invention

[0038] The present invention, configured as described above, has the advantage of being able to analyze the detailed lesion location for stroke, and in particular, to analyze the detailed lesion location for strokes such as small subcortical cerebral infarction or large vessel entrance occlusion stroke.

[0039] In addition, the present invention has the advantage of providing a visual guideline to clinicians for differentiating the causes of small subcortical cerebral infarction by providing a brain lesion template related to small subcortical cerebral infarction.

[0040] In addition, the present invention has the advantage of improving the accuracy of the analysis of small subcortical cerebral infarction by dividing each region of the infarction area to reflect the pathological cause of the small subcortical cerebral infarction that occurs due to the blockage of blood vessels in the subcortical brain region.

[0041] The effects obtainable from the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below. Brief explanation of the drawing

[0043] FIG. 1 shows a block diagram of a system (10) according to one embodiment of the present invention. Figure 2 shows examples of lesion locations when a small subcortical cerebral infarction occurs. FIG. 3 shows a block diagram of an analysis device (200) according to one embodiment of the present invention. FIG. 4 shows a block diagram of a control unit (250) in an analysis device (200) according to one embodiment of the present invention. FIG. 5 shows a flowchart of an analysis method according to one embodiment of the present invention. Figure 6 shows an example of S520 to S540. Figure 7 shows an example of the first-1 clustering image for the case where the number of clusters (k) is set to 12. Specific details for implementing the invention

[0044] The above-mentioned objectives, means, and resulting effects of the present invention will become clearer through the following detailed description in conjunction with the attached drawings, and accordingly, a person skilled in the art to which the present invention pertains will be able to easily implement the technical concept of the present invention. Furthermore, in describing the present invention, if it is determined that a detailed description of known technology related to the present invention may unnecessarily obscure the essence of the present invention, such detailed description will be omitted.

[0045] The terms used herein are for describing the embodiments and are not intended to limit the invention. In this specification, the singular form includes the plural form as appropriate unless specifically stated otherwise in the text. In this specification, terms such as "comprising," "providing," "making arrangements," or "having" do not exclude the presence or addition of one or more other components in addition to the components mentioned.

[0046] In this specification, terms such as "or", "at least one," etc., may indicate one of the words listed together or a combination of two or more. For example, "A or B", "at least one of A and B" may include only one of A or B, or may include both A and B.

[0047] In this specification, descriptions following "e.g." should not limit the embodiments of the invention according to various embodiments of the invention, such as variations including tolerances, measurement errors, limits of measurement accuracy, and other commonly known factors, as the information presented, such as cited characteristics, variables, or values, may not exactly match.

[0048] In this specification, where it is stated that a component is 'connected' or 'connected' to another component, it should be understood that it may be directly connected or connected to the other component, or that there may be other components in between. On the other hand, when it is mentioned that a component is 'directly connected' or 'directly connected' to another component, it should be understood that there are no other components in between.

[0049] In this specification, where a component is described as being 'on' or 'in contact' with another component, it should be understood that it may be in direct contact with or connected to the other component, but that another component may exist in between. On the other hand, where a component is described as being 'immediately above' or 'in direct contact' with another component, it should be understood that no other component exists in between. Other expressions describing the relationship between components, such as 'between' and 'directly between', may be interpreted in the same way.

[0050] In this specification, terms such as 'first,' 'second,' etc., may be used to describe various components, but such components should not be limited by these terms. Furthermore, these terms should not be interpreted as limiting the order of each component, but may be used for the purpose of distinguishing one component from another. For example, 'first component' may be named 'second component,' and similarly, 'second component' may be named 'first component.'

[0051] Unless otherwise defined, all terms used in this specification may be used in a meaning commonly understood by those skilled in the art to which the present invention pertains. Additionally, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0053] Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to the attached drawings.

[0054] FIG. 1 shows a block diagram of a system (10) according to one embodiment of the present invention.

[0055] A system (10) according to one embodiment of the present invention (hereinafter referred to as "the system") is a system for analyzing the detailed lesion location of a stroke, and is a system for analyzing the detailed lesion location of a small vessel occlusion stroke (a lacunar stroke in the traditional view), such as a small subcortical cerebral infarction, or a branch atheromatous stroke. Of course, the system may also be a system for assisting (i.e., providing assistance) the analysis by a user (medical professional, etc.).

[0056] At this time, a stroke is a disease in which brain cells receiving blood flow through blood vessels are damaged as a result of blood vessels supplying blood to the brain becoming blocked or rupturing. Furthermore, among these types of strokes, cerebral infarction, which occurs due to occlusion of blood vessels, is a disease in which a part of brain tissue dies because blood flow is interrupted due to a blockage of cerebral blood vessels.

[0057] Cerebral infarction can generally be classified into four categories based on clinical symptoms and the size of the lesion on brain magnetic resonance imaging. In particular, small subcortical infarction refers to a stroke occurring in small vessels within the brain; it corresponds to a cerebral infarction characterized by typical lacunar syndrome, the absence of cerebral cortical symptoms, and a lesion diameter of 20 mm or less (e.g., 1 mm to 20 mm). For instance, small subcortical infarction may include lacunar infarction. Such lacunar infarctions mostly occur in the periventricular white matter (PVWM) and basal ganglia (BG), while some occur in the cerebellum or brainstem. Lesions that appear similar to small vessel occlusion cerebral infarction on brain imaging may actually be large vessel inlet occlusion cerebral infarction; since treatment strategies and prognoses differ, differential diagnosis is important but has been a challenging issue. In addition, it can be classified into stroke caused by large artery atherosclerosis, which is a cerebral infarction in which a large cerebral blood vessel is stenotic or occluded by more than 50% due to atherosclerosis, there is cerebral cortex or cerebellar dysfunction, and the diameter of the lesion is greater than 20 mm, stroke caused by cardiometabolic encephalopathy, and stroke of unknown cause.

[0058] FIG. 2 shows examples of lesion locations when lacunar cerebral infarction occurs. Specifically, FIG. 2(a) shows an example of blockage of a small blood vessel (V1), and FIG. 2(b) shows an example of blockage of the entrance to a large blood vessel (V2).

[0059] In particular, in the case of a small subcortical cerebral infarction or a stroke with an occlusion of a large vessel opening, the treatment method and prognosis vary depending on the location of the occlusion (i.e., the location of the lesion) that causes the stroke. For example, referring to Fig. 2, the treatment method is applied differently depending on whether the cause of the small subcortical cerebral infarction is an occlusion of a small vessel (V1) corresponding to a small diameter among microvessels, or an occlusion of the opening of a large vessel (V2) corresponding to a superior vessel from which the small vessel (V1) branches.

[0060] However, even if a brain image of a patient with a small subcortical cerebral infarction is provided, the brain image does not provide detailed images of the small vessel (V1) occlusion or the large vessel opening (V2) occlusion as shown in FIG. 2, and only provides images of the stroke area of ​​the brain tissue damaged by the occlusion.

[0061] Accordingly, the present system (10) and the analysis method described below are intended to provide a technique for analyzing the detailed lesion location and cause of a stroke (particularly, a small subcortical cerebral infarction) in the medical image. For example, referring to FIG. 2, the present system (10) and the analysis method described below are intended to provide a technique for analyzing whether the small subcortical cerebral infarction is caused by occlusion of a small blood vessel (V1) or by occlusion of the entrance of a large blood vessel (V2).

[0062] This system (10) may include a medical imaging device (100) and an analysis device (200), as shown in FIG. 1.

[0063] The medical imaging device (100) is a device for taking medical images of a brain region of a patient with a brain disease who has suffered a stroke. That is, the medical image is an image taken by the medical imaging device (100) of a brain region of a patient with a small subcortical cerebral infarction, and may be a two-dimensional or three-dimensional image obtained using radiation, electromagnetic waves, etc. to identify the interior of the brain region.

[0064] For example, medical imaging may include computed tomography (CT) images, magnetic resonance imaging (MRI), etc. In particular, since the imaging concerns small subcortical cerebral infarctions such as lacunar cerebral infarction, the medical imaging may be MRI, and while diffusion-weighted MRI (DWI) may be preferable, it is not limited thereto.

[0065] In this context, DWI is image data obtained through a technique that measures the irregular movement of water particles within tissue, i.e., the diffusion rate; it is an image processed using a method generally capable of early detection of cerebral infarction or lesions caused by cerebral vascular occlusion. By detecting these signal changes, various types of ischemic brain tissue damage, such as the location and volume of damage, can be visualized. In this case, the diffusion rate is determined by the b-value, which represents the strength of the gradient magnetic field.

[0066] In particular, the medical imaging device (100) is connected to the analysis device (200) so that the captured medical image can be transmitted to the analysis device (200). Of course, the medical image may be separately stored in another server, device, or storage medium and then transmitted to the analysis device (200). For example, the medical image may be transmitted from the medical imaging device (100), another server or device, etc. to the analysis device (200) via an internal network or an external network such as the Internet, or stored in a storage medium such as a flash memory or an optical disc and then connected to the analysis device (200) so that it can be stored in the analysis device (200).

[0067] FIG. 3 shows a block diagram of an analysis device (200) according to one embodiment of the present invention.

[0068] The analysis device (200) is a device that processes medical images captured by the medical imaging device (100). That is, the analysis device (200) may be a device for analyzing the detailed lesion location and cause of a small subcortical cerebral infarction in a medical image, or for assisting (i.e., helping) the analysis by a user (medical professional, etc.). This analysis device (200) is a computing-capable electronic device installed in a medical field, etc., and may include an input unit (210), a communication unit (220), a display (230), a memory (140), and a control unit (250), as shown in FIG. 3.

[0069] For example, the electronic device may be a general-purpose computing system such as a desktop personal computer, laptop personal computer, tablet personal computer, netbook computer, workstation, personal digital assistant (PDA), smartphone, smartpad, or mobile phone, or a dedicated embedded system implemented based on Embedded Linux, but is not limited thereto.

[0070] The input unit (210) generates input data in response to input from various users and may include various input means.

[0071] For example, the input unit (210) may include a keyboard, key pad, dome switch, touch panel, touch key, touch pad, mouse, menu button, etc., but is not limited thereto.

[0072] The communication unit (220) is configured to perform communication with other devices, servers, or medical imaging devices (100), etc. (hereinafter referred to as "other devices, etc."). For example, the communication unit (220) can receive various data or medical images, etc., necessary for performing the analysis method described later from other devices, etc. Additionally, the communication unit (200) can transmit data regarding the results of the analysis method of the analysis device (200) described later to other devices, etc.

[0073] For example, the communication unit (220) can perform wireless communication such as 5G (5th generation communication), LTE-A (long term evolution-advanced), LTE (long term evolution), Bluetooth, BLE (Bluetooth low energy), NFC (near field communication), and WiFi communication, or wired communication such as cable communication, but is not limited thereto.

[0074] The display (230) displays various image data on a screen and may be composed of a non-emissive panel or an emissive panel. That is, the display (230) can display image data according to the performance of the analysis method described later.

[0075] For example, the display (230) may include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display, an organic light emitting diode (OLED) display, a micro electro mechanical systems (MEMS) display, or an electronic paper display. Additionally, the display (230) may be combined with an input unit (220, 220) to be implemented as a touch screen, etc.

[0076] The memory (240) stores various information necessary for the operation of the analysis device (200). The information stored in the memory (240) may include image information, information transmitted and received with other devices through the communication unit (220), information for the control operation of the control unit (250), and program information related to the analysis method to be described later, but is not limited thereto.

[0077] For example, depending on the type, the memory (240) may include a hard disk type, a magnetic media type, a CD-ROM (compact disc read only memory), an optical media type, a magneto-optical media type, a multimedia card micro type, a flash memory type, a ROM type (read only memory type), or a RAM type (random access memory type), but is not limited thereto. Additionally, depending on the use / location, the memory (240) may be a cache, a buffer, a main memory, or an auxiliary memory, or a separately provided storage system, but is not limited thereto.

[0078] The control unit (250) can perform various control operations of the analysis device (200). That is, the control unit (250) can perform control over the medical imaging device (100) or control the execution of the analysis method to be described later. For example, the control unit (250) can control the medical image to be displayed through the display (230), and can control the execution process and results of the analysis method to be described later to be displayed through the display (230).

[0079] Additionally, the control unit (250) can control the operation of the remaining components of the analysis device (200), namely the input unit (210), communication unit (220), display (230), memory (240), etc. Of course, the control unit (205) may also perform control over the medical imaging device (100). For example, the control unit (250) may include a processor, which is hardware, or a process, which is software executed on said processor, but is not limited thereto. For example, the processor may include a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), etc., but is not limited thereto.

[0080] FIG. 4 shows a block diagram of a control unit (250) in an analysis device (200) according to one embodiment of the present invention.

[0081] To control the execution of the analysis method described below, the control unit (250) may include a matching unit (251), a boundary setting unit (252), a dimensionality reduction unit (253), a clustering unit (254), a comparison unit (255), and an analysis unit (256), as shown in FIG. 4. For example, the matching unit (251), the boundary setting unit (252), the dimensionality reduction unit (253), the clustering unit (254), the comparison unit (255), and the analysis unit (256) may be hardware components of the control unit (250) or software processes executed by the control unit (250), but are not limited thereto.

[0082] Hereinafter, an analysis method according to one embodiment of the present invention will be described in more detail.

[0083] FIG. 5 shows a flowchart of an analysis method according to one embodiment of the present invention.

[0084] An analysis method according to one embodiment of the present invention (hereinafter referred to as "the analysis method") is a method performed under the control of a control unit (250) of an analysis device (200), and may include S510 to S560 as shown in FIG. 5. However, if necessary, at least one of S510 to S560 may be omitted. For example, S530 or S550 may be omitted.

[0085] However, when performing S510 to S560, medical images stored in memory (240) may be used. At this time, the medical images may be received from another device, etc., via the communication unit (220) through an internal network or an external network such as the Internet to the analysis device (200) and stored in memory (240), or stored in a storage medium such as a flash memory or an optical disc and then the storage medium is connected to the analysis device (200) so that they can be stored in memory (240).

[0086] These medical images can be divided into a first medical image used in S510 to S550 (i.e., a medical image for training) and a second medical image (i.e., a target medical image) to which the results of S510 to S550 are applied. It is preferable that these first and second medical images are medical images of different patients.

[0087] In this case, the first medical image is a medical image of the brain region of a patient who was previously treated and had a small subcortical infarction. On the other hand, the second medical image is a medical image of the brain region of a patient who is currently treated and is predicted to have or may have a small subcortical infarction.

[0088] The present analysis method relates to a technique for analyzing the detailed lesion location and cause of a small subcortical cerebral infarction in a second medical image by performing S510 to S550 based on a number of first medical images of patients who were previously treated, and then applying a second medical image of a patient currently being treated to the results of S510 to S550.

[0089] These S510 to S550 steps can be performed in an analysis device (200) as described below. At this time, the control unit (250) may control the transmission of the results of S510 to S550 to another device, which is a computing-capable electronic device, through the communication unit (220). In this case, the analysis device (200) may function as a server that provides the results of S510 to S550 to the other device.

[0090] Additionally, the analysis device (200) may perform only S560. In this case, S510 to S550 may be performed on another device or server capable of computing, as described below. At this time, the results of S510 to S550 performed on another device or server may be received from the other device or server via the communication unit (220) to the analysis device (200) via an internal network or an external network such as the Internet and stored in memory (240), or stored on a storage medium such as flash memory or an optical disc, and then the storage medium may be connected to the analysis device (200) to be stored in memory (240).

[0091] Additionally, the analysis device (200) stores the results of S510 to S550, receives a second medical image from another device capable of computing, performs S560 on the received second medical image, and may transmit the result of S560 to the other device. In this case, the analysis device (200) may function as a server that provides the result of S560 to the other device.

[0092] Meanwhile, the first and second medical images used in S510 to S560 may be images that include information regarding the stroke site. That is, the first and second medical images include image data taken of the brain area of ​​a stroke patient and data regarding the location of the area where the stroke occurred within that brain area (i.e., the location of the lesion).

[0093] Hereinafter, the occlusion of the small blood vessel (V1) is referred to as "first occlusion," and the occlusion of the entrance of the large blood vessel (V2) is referred to as "second occlusion."

[0094] Additionally, the first medical image can be classified into the first-1 medical image and the first-2 medical image. In this case, the first-1 medical image is the first medical image of a past patient subject to treatment who developed a small subcortical cerebral infarction due to an occlusion of either the first or second occlusion. Conversely, the first-2 medical image is the first medical image of a past patient subject to treatment who developed a small subcortical cerebral infarction due to an occlusion of the other of the first or second occlusions. Of course, the second medical image is the second medical image of a current patient subject to treatment who has developed, or is predicted to develop, a small subcortical cerebral infarction due to the first or second occlusion.

[0095] Below, the detailed contents of S510 to S550 will be explained.

[0096] <1st Example>

[0097] First, in S510, the registration unit (251) registers a plurality of first medical images. At this time, each first medical image is an image obtained from different patients and, accordingly, includes images of different brain shapes. Accordingly, by registering each first medical image, the brain shapes of each first medical image can be registered and become similar.

[0098] For example, multiple first medical images can be aligned based on a standard medical image. In this case, the standard medical image is a medical image taken of the corresponding brain region, and is an image that includes a standard brain shape for each medical image, as the shape of the brain inevitably differs from person to person. Accordingly, the alignment unit (251) aligns the different brain shapes of different patients included in each first medical image with the standard brain shape included in the standard medical image. As a result, the brain shape of each aligned first medical image can become similar to the brain shape of the standard medical image. Of course, as a result of the alignment process, data regarding the location of the stroke area included in each first medical image is also corrected.

[0099] For such registration processing, various conventional image registration algorithms for registering heterogeneous images may be used. For example, for registration processing of multiple first medical images, linear and nonlinear rigid body registration algorithms may be used, but are not limited thereto.

[0100] For example, there may be 1,000 first medical images, of which 600 are first-1 medical images and the remaining 400 are first-2 medical images.

[0101] FIG. 6 illustrates an example of S520 to S540. Specifically, FIG. 6(a) illustrates an example of the execution process of S520, FIG. 6(b) illustrates an example of the execution process of S530, FIG. 6(c) illustrates an example of the execution process of S540, and FIG. 6(d) illustrates an example of the result of the execution of S540.

[0102] Next, in S520, the boundary setting unit (252) sets a common area where a small subcortical cerebral infarction has occurred among a plurality of aligned medical images (1-1), and sets a boundary (i.e., a brain lesion boundary) for the common area.

[0103] That is, the boundary setting unit (252) can set spatially connected brain lesion boundaries by extracting voxels (hereinafter referred to as "common voxels") of overlapping lesion locations from at least N (where N is a natural number greater than or equal to 2 or 3) of the matched 1-1 medical images, and by setting a common region including these common voxels and setting a boundary for the common region. At this time, a voxel refers to a normal grid unit in the space of a 3-dimensional image, and may refer to a pixel in a 2-dimensional plane image.

[0104] In particular, it may be preferable that the common voxel is a voxel of a lesion location that overlaps in at least two different medical images. However, for higher accuracy of the lesion location, it may be more preferable that the common voxel is a voxel of a lesion location that overlaps in at least three different medical images.

[0105] For example, as illustrated in FIG. 6(a), the boundary setting unit (252) may mark the voxels of the common area corresponding to the common voxels that overlap with the three first-1 medical images in a three-dimensional area corresponding to the brain region of a standard medical image as 1, and mark the voxels of the area corresponding to the remaining voxels other than the common voxels as 0. That is, the boundary setting unit (252) may set the boundary of the area marked as 1 in the corresponding three-dimensional area as the brain lesion boundary.

[0106] Next, in S530, the dimension reduction unit (253) reduces the dimension of the data of the common voxel (i.e., the data of the common area) according to the setting of the brain lesion boundary to remove unnecessary information.

[0107] For example, as illustrated in FIG. 6(b), the dimensionality reduction unit (253) can reduce the dimensionality of the data of the common voxel (i.e., the data of the common region) by performing logistic non-negative matrix factorization (logistic NMF). Data with dimensions of the number of voxels x the number of patients is transformed into data with dimensions of the number of voxels x the number of reduced dimensions through dimensionality reduction in the patient dimension. This process compresses information about the patient, thereby preventing data overfitting during the subsequent data clustering of S540. However, this S530 may be omitted.

[0108] Next, in S540, the clustering unit (254) performs clustering on the data of the common voxels to divide the common region into multiple regions. At this time, each divided region contains information on the co-occurrence of similar lesions within the brain lesion boundary. Of course, when S530 is performed, the data of the common voxels to be clustered may be dimensionally reduced data.

[0109] For example, as illustrated in FIG. 6(c), the clustering unit (254) can divide multiple regions by performing hierarchical clustering in the spatial dimension on the data of common voxels in multiple first-1 medical images. At this time, the number of clusters (k) can be varied as needed. That is, k is a natural number greater than or equal to 2.

[0110] Figure 7 shows an example of the first-1 clustering image for the case where the number of clusters (k) is set to 12.

[0111] As a result of performing S540, as the data of common voxels in multiple first-1 medical images are clustered, an image (hereinafter referred to as the "first-1 clustered image") can be derived in which multiple clustered (i.e., segmented) regions are displayed in a standard medical image. At this time, as illustrated in FIGS. 6(d) and FIGS. 7, each segmented region may be displayed in a different color in the first-1 clustered image.

[0112] Next, in S550, the comparison unit (255) tests the suitability of the 1-1 clustered image by comparing the stroke area of ​​the 1-2 medical image with each clustered region of the 1-1 clustered image. At this time, it is preferable that the 1-2 medical image be an image that has been aligned with a standard medical image according to the process described above in S510.

[0113] That is, the comparison unit (255) can compare the first-1 clustered image with the first-2 medical image one by one. At this time, since the first-1 clustered image is an image in which data of common voxels from multiple first-1 medical images are clustered, the detailed location and cause of the stroke lesion differ from that of the first-2 medical image. Accordingly, when making the comparison, a significant difference is bound to occur between the stroke area of ​​the first-2 medical image and each clustered area of ​​the first-1 clustered image.

[0114] For example, the first medical image may be a first medical image of a small subcortical cerebral infarction caused by a first blockage (i.e., blockage of a small blood vessel). In this case, the stroke area of ​​the first medical image caused by a second blockage (i.e., blockage of the entrance of a large blood vessel) generally occupies a larger area than the clustered regions of the first clustered image. At this time, if the area of ​​the stroke area of ​​the first medical image that extends beyond the clustered regions of the first clustered image (i.e., the difference area between the stroke area of ​​the first medical image and the clustered regions of the first clustered image) is greater than a certain amount, the comparison unit (255) may determine that the first clustered image is suitable. On the other hand, if the difference area is less than a certain amount, the comparison unit (255) may determine that the first clustered image is unsuitable.

[0115] Meanwhile, the 1-1 medical image may be the 1st medical image of a small subcortical cerebral infarction caused by the 2nd blockage (i.e., blockage of the entrance of a large vessel). In this case, the stroke area of ​​the 1-2 medical image caused by the 1st blockage (i.e., blockage of a small vessel) generally occupies a narrower area than each clustered region of the 1-1 clustered image. At this time, if the area outside the stroke area of ​​the 1-2 medical image among each clustered region of the 1-1 clustered image (i.e., the difference area between each clustered region of the 1-1 clustered image and the stroke area of ​​the 1-2 medical image) is greater than a certain amount, the comparison unit (255) may determine that the 1-1 clustered image is suitable. On the other hand, if the difference area is less than a certain amount, the comparison unit (255) may determine that the 1-1 clustered image is unsuitable.

[0116] Next, in S560, the analysis unit (256) provides a template of a first-1 clustered image (also referred to as a "brain lesion template") containing each clustered region. At this time, the template of the first-1 clustered image may be a template for the first-1 clustered image determined to be suitable in S550.

[0117] That is, the analysis unit (256) can display the templates of the second medical image and the first-1 clustered image on the display (230), respectively. At this time, the template of the first-1 clustered image can be displayed so as to overlap the second medical image. That is, the clustered regions of the template of the first-1 clustered image and the stroke areas of the second medical image can be displayed so as to overlap. Through this, a user (e.g., a medical professional) can easily analyze the detailed lesion location and cause of the small subcortical cerebral infarction in the second medical image.

[0118] For example, if the stroke area in the second medical image is similar to each clustered region in the first-1 clustered image, the user can analyze that the detailed lesion location and cause of the stroke in the second medical image are the same as the detailed lesion location and cause of the stroke in the first-1 clustered image (i.e., analyzed as being the same as the detailed lesion location and cause of the stroke in the first-1 medical image).

[0119] Of course, the analysis unit (256) can directly analyze the detailed lesion location and cause of the small subcortical cerebral infarction in the second medical image by comparing the template of the first-1 clustered image with the stroke area of ​​the second medical image.

[0120] That is, the analysis unit (256) can compare the similarity between each clustered region of the first-1 clustered image and the stroke region of the second medical image. At this time, since the first-1 clustered image is an image in which data of common voxels from multiple first-1 medical images are clustered, if, as a result of the comparison, the stroke region of the second medical image is similar to each clustered region of the first-1 clustered image (i.e., if the similarity is above a certain level), the analysis unit (256) can analyze that the detailed lesion location and cause of the stroke in the second medical image is the same as the detailed lesion location and cause of the stroke in the first-1 clustered image (i.e., analyzed as being the same as the detailed lesion location and cause of the stroke in the first-1 medical image).

[0121] On the other hand, if, as a result of the comparison, the stroke area of ​​the second medical image is not similar to each clustered region of the first-1 clustered image (i.e., if the similarity is less than a certain level), the analysis unit (256) can analyze that the detailed lesion location and cause of the stroke in the second medical image is different from the detailed lesion location and cause of the stroke in the first-1 clustered image (i.e., analyzed as being the same as the detailed lesion location and cause of the stroke in the first-2 medical image).

[0122] For example, the first medical image may be a first medical image of a small subcortical cerebral infarction caused by a first blockage (i.e., a blockage of a small blood vessel). In this case, if the difference between the stroke area of ​​the second medical image and each clustered region of the first-1 clustered image is greater than a certain amount (i.e., if the similarity is less than or greater than a certain amount), the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small subcortical cerebral infarction occurred due to a second blockage of a different type from the first blockage. On the other hand, if the difference is less than a certain amount (i.e., if the similarity is greater than or equal to a certain amount), the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small blood vessel subcortical cerebral infarction occurred due to a first blockage.

[0123] Meanwhile, the first medical image may be a first medical image of a small subcortical cerebral infarction caused by a second blockage (i.e., a blockage of the entrance of a major blood vessel). In this case, if the difference between the stroke area of ​​the second medical image and each clustered area of ​​the first-1 clustered image is greater than a certain amount, the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small subcortical cerebral infarction occurred due to a first blockage of a different type from the second blockage. On the other hand, if the difference is less than a certain amount, the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small subcortical cerebral infarction occurred due to a second blockage.

[0124] <2nd Example>

[0125] In the case of the second embodiment, some details may be changed from the first embodiment described above. Hereinafter, only these changed parts will be described.

[0126] First, S510 is performed identically.

[0127] Next, in S520, the boundary setting unit (252) sets a first common area in which a small subcortical cerebral infarction has occurred among a plurality of aligned medical images (1-1), and sets a brain lesion boundary (i.e., a first brain lesion boundary) for the said first common area. Additionally, the boundary setting unit (252) sets a second common area in which a small subcortical cerebral infarction has occurred among a plurality of aligned medical images (1-2), and sets a brain lesion boundary (i.e., a second brain lesion boundary) for the said second common area.

[0128] That is, the boundary setting unit (252) can set a spatially connected first brain lesion boundary by extracting first common voxels of overlapping lesion locations in at least N (wherein N is a natural number greater than or equal to 2 or 3) of the matched first-1 medical images, and by setting a first common region including these first common voxels and setting a boundary for the first common region.

[0129] Additionally, the boundary setting unit (252) can set spatially connected second brain lesion boundaries by extracting second common voxels of overlapping lesion locations from at least N (wherein N is a natural number greater than or equal to 2 or 3) of the matched first and second medical images, and by setting a second common region including these second common voxels and setting a boundary for the second common region.

[0130] Next, in S530, the dimension reduction unit (253) reduces the dimension of the data of the first common voxel (i.e., the data of the first common region) according to the setting of the first brain lesion boundary, and reduces the dimension of the data of the second common voxel (i.e., the data of the second common region) according to the setting of the second brain lesion boundary.

[0131] Next, in S540, the clustering unit (254) performs clustering on the data of the first common voxel to divide the first common region according to the first brain lesion boundary into multiple regions. Additionally, the clustering unit (254) performs clustering on the data of the second common voxel according to the setting of the second brain lesion boundary to divide the second common region according to the second brain lesion boundary into multiple regions.

[0132] As a result of performing S540, the aforementioned 1-1 clustered image and 1-2 clustered image can be derived, respectively. In this case, the 1-2 clustered image is an image in which multiple clustered (i.e., segmented) regions are displayed in a standard medical image as the data of common voxels in multiple 1-2 medical images are clustered. In this case, each segmented region may be displayed in a different color in the 1-1 and 1-2 clustered images, respectively.

[0133] Next, in S550, the comparison unit (255) tests the suitability of the 1-1 clustered image by comparing the stroke area of ​​the 1-2 medical image with each clustered region of the 1-1 clustered image. At this time, the comparison of the 1-2 medical image with the 1-1 clustered image is as described above in S540.

[0134] Additionally, the comparison unit (255) tests the suitability of the 1-2 clustered image by comparing the stroke area of ​​the 1-1 medical image with each clustered area of ​​the 1-2 clustered image.

[0135] That is, the comparison unit (255) can compare the 1-1 medical image with the 1-2 clustered image one by one. At this time, since the 1-2 clustered image is an image in which data of common voxels from multiple 1-2 medical images are clustered, the detailed location and cause of the stroke lesion differ from the 1-1 medical image. Accordingly, when comparing, a significant difference is bound to occur between the stroke area of ​​the 1-1 medical image and each clustered area of ​​the 1-2 clustered image.

[0136] For example, the first-2 medical image may be a first medical image of a small subcortical cerebral infarction caused by a first blockage (i.e., blockage of a small vessel). In this case, the stroke area of ​​the first-1 medical image caused by a second blockage (i.e., blockage of the entrance of a large vessel) generally occupies a larger area than the clustered regions of the first-2 clustered image. At this time, if the area of ​​the stroke area of ​​the first-1 medical image that extends beyond the clustered regions of the first-2 clustered image (i.e., the difference area between the stroke area of ​​the first-1 medical image and the clustered regions of the first-2 clustered image) is greater than a certain amount, the comparison unit (255) may determine that the first-2 clustered image is suitable. On the other hand, if the difference area is less than a certain amount, the comparison unit (255) may determine that the first-2 clustered image is unsuitable.

[0137] Meanwhile, the first-2 medical image may be the first medical image of a small subcortical cerebral infarction caused by the second blockage (i.e., blockage of the entrance of a large vessel). In this case, the stroke area of ​​the first-1 medical image caused by the first blockage (i.e., blockage of a small vessel) generally occupies a narrower area than each clustered region of the first-2 clustered image. At this time, if the area outside the stroke area of ​​the first-1 medical image among each clustered region of the first-2 clustered image (i.e., the difference area between each clustered region of the first-2 clustered image and the stroke area of ​​the first-1 medical image) is greater than a certain amount, the comparison unit (255) may determine that the first-2 clustered image is suitable. On the other hand, if the difference area is less than a certain amount, the comparison unit (255) may determine that the first-2 clustered image is unsuitable.

[0138] Next, in S560, the analysis unit (256) provides a template for the first-1 and first-2 clustered images (also referred to as "brain lesion template") each containing a clustered region. At this time, the template for the first-1 and first-2 clustered images may be a template for the first-1 and first-2 clustered images determined to be suitable in S550.

[0139] That is, the analysis unit (256) can display the second medical image and the templates of the first-1 and first-2 clustered images, respectively, on the display (230). At this time, the template of the first-1 clustered image can be displayed so as to overlap the second medical image, and the template of the first-2 clustered image can be displayed so as to overlap the second medical image. That is, the clustered areas of the template of the first-1 clustered image and the stroke areas of the second medical image can be displayed so as to overlap, and the clustered areas of the template of the first-2 clustered image and the stroke areas of the second medical image can be displayed so as to overlap. Through this, a user (e.g., a medical professional) can easily analyze the detailed lesion location and cause of the small subcortical cerebral infarction in the second medical image.

[0140] For example, if the stroke area in the second medical image is more similar to each clustered region of the first-1 clustered image, the user can analyze that the detailed lesion location and cause of the stroke in the second medical image are identical to the detailed lesion location and cause of the stroke in the first-1 clustered image (i.e., analyzed as identical to the detailed lesion location and cause of the stroke in the first-1 medical image). On the other hand, if the stroke area in the second medical image is more similar to each clustered region of the first-2 clustered image, the user can analyze that the detailed lesion location and cause of the stroke in the second medical image are identical to the detailed lesion location and cause of the stroke in the first-2 clustered image (i.e., analyzed as identical to the detailed lesion location and cause of the stroke in the first-2 medical image).

[0141] Of course, the analysis unit (256) can compare the similarity between the template of the first-1 clustered image and the stroke area of ​​the second medical image (hereinafter referred to as "first comparison") and compare the similarity between the template of the first-2 clustered image and the stroke area of ​​the second medical image (hereinafter referred to as "second comparison") to directly analyze the detailed lesion location and cause of the small subcortical cerebral infarction of the second medical image.

[0142] At this time, if, as a result of the comparison between the first and second medical images, the stroke area of ​​the second medical image is more similar to each clustered region of the first-1 clustered image, the analysis unit (256) can analyze that the detailed lesion location and cause of the stroke in the second medical image is the same as the detailed lesion location and cause of the stroke in the first-1 clustered image (i.e., analyze that it is the same as the detailed lesion location and cause of the stroke in the first-1 medical image).

[0143] On the other hand, if the stroke area of ​​the second medical image is more similar to each clustered area of ​​the first-2 clustered image, the analysis unit (256) can analyze that the detailed lesion location and cause of the stroke in the second medical image is the same as the detailed lesion location and cause of the stroke in the first-2 clustered image (i.e., analyzed as being the same as the detailed lesion location and cause of the stroke in the first-2 medical image).

[0144] For example, the first-1 medical image may be a first medical image of a small subcortical cerebral infarction caused by a first blockage (i.e., blockage of a small blood vessel), and the first-2 medical image may be a first medical image of a small subcortical cerebral infarction caused by a second blockage (i.e., blockage of the entrance of a large blood vessel). In this case, if the similarity value between the stroke area of ​​the second medical image and each clustered region of the first-1 clustered image (hereinafter referred to as "first similarity") is greater than the similarity value between the stroke area of ​​the second medical image and each clustered region of the first-2 clustered image (hereinafter referred to as "second similarity"), the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small subcortical cerebral infarction occurred due to a first blockage. On the other hand, if the second similarity is greater than the first similarity, the analysis unit (256) can analyze that the second medical image is a medical image of a patient who has developed a small subcortical cerebral infarction due to the second blockage.

[0145] Meanwhile, the 1-1 medical image may be a first medical image of a small subcortical cerebral infarction caused by a second blockage (i.e., blockage of the entrance of a large blood vessel), and the 1-2 medical image may be a first medical image of a small subcortical cerebral infarction caused by a first blockage (i.e., blockage of a small blood vessel). In this case, if the first similarity is greater than the second similarity, the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small subcortical cerebral infarction occurred due to the second blockage. On the other hand, if the second similarity is greater than the first similarity, the analysis unit (256) may analyze that the second medical image is a medical image of a patient in whom a small subcortical cerebral infarction occurred due to the first blockage.

[0147] The present invention, configured as described above, has the advantage of being able to analyze the detailed location of lesions in strokes, and in particular, to analyze the detailed location of lesions in small subcortical cerebral infarctions such as lacunar cerebral infarctions. Furthermore, the present invention has the advantage of providing visual guidelines to clinicians for the diagnosis of small subcortical cerebral infarctions by providing a brain lesion template related to small subcortical cerebral infarctions. Additionally, the present invention has the advantage of improving the accuracy of the analysis of small subcortical cerebral infarctions by dividing each region of the stroke site to reflect the pathological cause of small subcortical cerebral infarctions that occur due to the blockage of blood vessels in the subcortical region.

[0149] Although specific embodiments have been described in the detailed description of the present invention, it is understood that various modifications are possible within the scope of the invention. Therefore, the scope of the present invention is not limited to the described embodiments and should be defined by the claims set forth below and equivalents thereof. Explanation of the symbols

[0151] 10: System 100: Medical Imaging Device 200: Analysis device 210: Input section 220: Communications Unit 230: Display 240: Memory 250: Control unit 251: Matching section 252: Boundary setting section 253: Dimensionality Reduction Section 254: Clustering Section 255: Comparison Section 256: Analysis Section

Claims

Claim 1 A method performed on an electronic device to analyze whether a vascular occlusive stroke is caused by a first blockage corresponding to a blockage of a small vessel corresponding to a small diameter vessel among microvessels, or by a second blockage corresponding to a blockage of the entrance of a large vessel corresponding to a superior vessel from which the said small vessel differentiates, comprising: a step of aligning a plurality of first medical images of brain regions of vascular occlusive stroke patients so that the brain shape is aligned with said plurality of first medical images; a step of establishing a common region including common voxels, which are voxels of overlapping lesion locations in at least N (where N is a natural number greater than or equal to 2) of the aligned first medical images, and establishing a spatially connected brain lesion boundary by establishing a brain lesion boundary for said common region; a step of deriving a clustered image, which is an image in which each divided region is displayed, by performing clustering on the data of said common region to divide said common region according to the brain lesion boundary; and a template for said clustered image and the brain of a vascular occlusive stroke patient currently being treated A method comprising: a step of displaying a second medical image of a region, wherein the clustered region of the template and the stroke region of the second medical image overlap; and a step of determining that the vascular occlusive stroke of the patient in the second medical image was caused by the first blockage or the second blockage based on the similarity between the clustered region of the clustered image and the stroke region of the second medical image; wherein the plurality of first medical images correspond to medical images of different past patients in which a vascular occlusive stroke occurred due to either the first or the second blockage. Claim 2 A method according to claim 1, wherein between the setting step and the deriving step, the method further includes a step of reducing the dimensionality of the data in the common area, and the deriving step includes a step of performing clustering on the data in the common area with reduced data dimensionality. Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 A method according to claim 1, wherein the determining step comprises determining that the patient in the second medical image has suffered a vascular occlusion stroke caused by a first blockage when the similarity is above a certain level, and determining that the patient in the second medical image has suffered a vascular occlusion stroke caused by a second blockage when the similarity is below a certain level. Claim 8 A method according to claim 1, wherein the plurality of first medical images each include a first-1 medical image of patients who have developed a vascular occlusion stroke due to a blockage of either the first or second blockage, and a first-2 medical image of patients who have developed a vascular occlusion stroke due to a blockage of the other of the first and second blockages. Claim 9 In claim 8, the plurality of first medical images used in the setting step are the first-1 medical images, and after the deriving step, the method further includes a step of testing the suitability of the clustered image by comparing the clustered image with the first-2 medical image, wherein the testing step determines that the clustered image is suitable if the difference between the stroke area of ​​the first-2 medical image and each clustered region of the clustered image is greater than a certain level, and determines that the clustered image is unsuitable if the difference is less than a certain level. Claim 10 delete Claim 11 In claim 8, the setting step includes the step of setting a first common area which is the common area for a plurality of first-1 medical images and the step of setting a second common area which is the common area for a plurality of first-2 medical images, and the deriving step includes the step of deriving a first clustered image which is the clustered image for the data of the first common area and deriving a second clustered image which is the clustered image for the data of the second common area. Claim 12 A memory storing multiple first medical images of brain regions of patients with vascular occlusion stroke; and a control unit that analyzes, using the plurality of first medical images, whether the vascular occlusive stroke of a patient currently being treated is caused by a first blockage corresponding to a blockage of a small vessel corresponding to a small diameter among microvessels, or by a second blockage corresponding to a blockage of the entrance of a large vessel corresponding to a superior vessel where the small vessel differentiates; wherein the control unit performs alignment processing of the plurality of first medical images so that the brain shape of each first medical image is aligned, sets a common region including common voxels which are voxels of overlapping lesion locations in at least N (where N is a natural number greater than or equal to 2) of the aligned first medical images, sets a spatially connected brain lesion boundary by setting a brain lesion boundary for the common region, and performs clustering on the data of the common region to divide the common region according to the brain lesion boundary into a plurality of regions, thereby deriving a clustered image in which each divided region is displayed, and A device that displays a template for clustered images and a second medical image of a brain region of a patient with a vascular occlusion stroke who is currently the subject of treatment, wherein the clustered regions of the template and the stroke region of the second medical image are displayed so as to overlap, and determines that the vascular occlusion stroke of the patient in the second medical image was caused by the first blockage or the second blockage based on the similarity between the clustered regions of the clustered images and the stroke region of the second medical image, wherein the plurality of first medical images correspond to medical images of different past patients in which a vascular occlusion stroke occurred due to either the first or the second blockage. Claim 13 In claim 12, the control unit controls to reduce the dimension of the data in the common area and controls to perform clustering on the data in the common area with reduced data dimension. Claim 14 delete Claim 15 delete Claim 16 A device according to claim 12, wherein the control unit determines that the patient in the second medical image has suffered a vascular occlusion stroke caused by a first blockage when the similarity is above a certain level, and determines that the patient in the second medical image has suffered a vascular occlusion stroke caused by a second blockage when the similarity is below a certain level. Claim 17 In claim 12, the first medical image comprises a first-1 medical image for patients who have suffered a vascular occlusion stroke due to a blockage of either the first or second blockage, and a first-2 medical image for patients who have suffered a vascular occlusion stroke due to a blockage of the other of the first and second blockages, respectively; the control unit controls the setting of the common area using a plurality of first-1 medical images, tests the suitability of the clustered image by comparing the clustered image with the first-2 medical image, and during the test, determines that the clustered image is suitable if the difference between the stroke area of ​​the first-2 medical image and each clustered area of ​​the clustered image is greater than a certain level, and determines that the clustered image is unsuitable if the difference is less than a certain level. Claim 18 delete Claim 19 In claim 17, the control unit controls to set a first common area, which is a common area for a plurality of first-1 medical images, and to set a second common area, which is a common area for a plurality of first-2 medical images; controls to derive a first clustered image, which is a clustered image for the data of the first common area, and to derive a second clustered image, which is a clustered image for the data of the second common area; and controls to display the first and second clustered images and a second medical image, which is a brain region of a patient with vascular occlusion stroke who is the subject of treatment, on a display, respectively. Claim 20 delete

Images