Search

KR-102963253-B1 - ELECTRONIC APPARTUS, CONTROL METHOD, SYSTEM AND COMPUTER PROGRAM FOR DIAGNOSING CEREBRAL INFARCTION BASED ON CT IMAGE OF SUBJECT PATIENT

KR102963253B1KR 102963253 B1KR102963253 B1KR 102963253B1KR-102963253-B1

Abstract

An electronic device is disclosed. The electronic device according to the present disclosure comprises a memory including a first artificial intelligence model trained to perform a diagnosis of cerebral infarction, and a processor that inputs an image of a subject patient to the first artificial intelligence model and obtains diagnostic information regarding the subject patient's cerebral infarction according to the output of the first artificial intelligence model, wherein the image is characterized in that a brain region of the subject patient is captured through computed tomography (CT) scanning.

Inventors

  • 김재국
  • 강구현
  • 장용수
  • 김원희
  • 최현영
  • 이윤재
  • 이가영
  • 김빛나래

Assignees

  • 한림대학교 산학협력단

Dates

Publication Date
20260508
Application Date
20240123

Claims (9)

  1. In electronic devices, A memory comprising a first artificial intelligence model trained to perform a diagnosis of cerebral infarction; and A processor that inputs an image of a target patient into the first artificial intelligence model and obtains diagnostic information regarding the target patient's cerebral infarction according to the output of the first artificial intelligence model; The image captured above is, Characterized by the fact that the brain region of the subject patient was imaged through computed tomography (CT), The above-mentioned first artificial intelligence model is, Identify at least one low-density region in an image of the brain region of the input target patient, and The above processor is, Identifying the identified low-density area as the area where the cerebral infarction occurred, The above-mentioned first artificial intelligence model is, Multiple patients with specified onset times of cerebral infarction are trained to output the progression pattern of cerebral infarction based on changes in low-density areas, based on multiple images taken at different times. The above processor is, Diagnostic information including the time at which the cerebral infarction of the subject patient occurred is obtained through the output of the first artificial intelligence model into which an image of the subject patient is input, and The above memory is, It further includes a second artificial intelligence model trained to output the probability of administering a thrombolytic agent for the treatment of cerebral infarction, and The above processor is, An electronic device that inputs the region where the cerebral infarction occurred, the time of occurrence of the cerebral infarction, and patient information for the target patient into the second artificial intelligence model to obtain the probability of administering a thrombolytic agent to the target patient.
  2. In claim 1, The above-mentioned first artificial intelligence model is, An electronic device characterized by being trained based on images of multiple normal people without cerebral infarction and images of multiple patients with specified onset times of cerebral infarction.
  3. delete
  4. delete
  5. delete
  6. In claim 1, The above processor is, Based on the output of the first artificial intelligence model, a first low-density region is identified within the image of the target patient, and If the subject patient has a history of past cerebral infarction, a second low-density region is identified within the image of the subject patient that matches the history of past cerebral infarction, and An electronic device that identifies the area in which the second low-density area is excluded from the first low-density area as the area where a cerebral infarction has occurred.
  7. In a system for diagnosing cerebral infarction in a target patient, A computed tomography (CT) scanner that photographs the brain portion of the above-mentioned patient and provides it to an electronic device; and The electronic device for diagnosing a cerebral infarction in a patient based on an image of the brain portion of the patient; comprising The above electronic device is, It includes a memory comprising a first artificial intelligence model trained to perform a diagnosis of cerebral infarction, and a processor that inputs an image of the subject patient into the first artificial intelligence model and obtains diagnostic information regarding the subject patient's cerebral infarction according to the output of the first artificial intelligence model. The above-mentioned first artificial intelligence model is, Identify at least one low-density region in an image of the brain portion of the input subject patient, and The above processor is, Identifying the identified low-density area as the area where the cerebral infarction occurred, The above-mentioned first artificial intelligence model is, Multiple patients with specified onset times of cerebral infarction are trained to output the progression pattern of cerebral infarction based on changes in low-density areas, based on multiple images taken at different times. The above processor is, Diagnostic information including the time at which the cerebral infarction of the subject patient occurred is obtained through the output of the first artificial intelligence model into which an image of the subject patient's brain portion is input, and The above memory is, It further includes a second artificial intelligence model trained to output the probability of administering a thrombolytic agent for the treatment of cerebral infarction, and The above processor is, A system for diagnosing cerebral infarction, which inputs the area where the cerebral infarction occurred, the time of occurrence of the cerebral infarction, and patient information into the second artificial intelligence model to obtain the probability of administering a thrombolytic agent to the target patient.
  8. In the method of operating an electronic device, The above electronic device inputs an image of a brain portion of a target patient into a first artificial intelligence model that has been trained to perform a diagnosis of at least one cerebral infarction; The electronic device comprises the step of obtaining an output regarding the cerebral infarction of the target patient from the first artificial intelligence model; and The electronic device comprises the step of determining the presence or absence of cerebral infarction in the target patient based on the output of the first artificial intelligence model; The above-mentioned first artificial intelligence model is, It is characterized by being trained to output the progression pattern of cerebral infarction according to changes in low-density areas based on multiple images taken at different times of multiple patients with specified times of onset of cerebral infarction, Identifying at least one low-density region in an image of the brain portion of the target patient input into the first artificial intelligence model, and The step of determining the presence or absence of cerebral infarction in the above-mentioned patient is, The above-mentioned identified low-density area is determined to be the area where a cerebral infarction occurred, and The method of operation of the above electronic device is, The electronic device comprises the step of obtaining diagnostic information including the time at which the cerebral infarction of the target patient occurred through the output of the first artificial intelligence model; and A method of operation of an electronic device comprising the step of: inputting the area where the cerebral infarction occurred for the target patient, the time of occurrence of the cerebral infarction, and patient information into a second artificial intelligence model trained to output the probability of administering a thrombolytic agent for the treatment of the cerebral infarction, thereby obtaining the probability of administering a thrombolytic agent to the target patient.
  9. A non-transient computer-readable medium storing at least one instruction that is executed by a processor of an electronic device to cause the electronic device to perform the method of operation of claim 8.

Description

ELECTRONIC APPARTUS, CONTROL METHOD, SYSTEM AND COMPUTER PROGRAM FOR DIAGNOSING CEREBRAL INFARCTION BASED ON CT IMAGE OF SUBJECT PATIENT } The present disclosure relates to an electronic device for diagnosing cerebral infarction, and more specifically, to an electronic device for calculating diagnostic information related to whether or not a cerebral infarction has occurred in a patient and the timing of the occurrence of the cerebral infarction based on an image of the patient's brain portion captured by computed tomography, or for determining whether to administer a thrombolytic agent to a patient identified as having cerebral infarction. According to statistics from the World Stroke Organization, stroke is the second leading cause of death worldwide and ranks first among single-organ diseases, excluding cancer which manifests in various sites. Stroke refers to neurological abnormalities such as unilateral paralysis, speech disorders, and impaired consciousness that occur when blood vessels supplying the brain become blocked (ischemic stroke) or rupture (hemorrhagic stroke), causing damage to the brain. The causes of stroke are diverse, including hypertension, diabetes, arrhythmias, and heart diseases such as arteriosclerosis. At this time, cerebral infarction (ischemic stroke) caused by a blockage of blood vessels accounts for the majority of all strokes, and acute cerebral infarction is the most common type, in which a cerebral blood vessel is suddenly blocked and blood flow supply decreases, causing brain cells to die. Cerebral infarction can be broadly classified into cerebral thrombosis, in which blood coagulates and forms a blood clot as arteriosclerosis develops in the cerebral blood vessels and causes them to gradually narrow, eventually growing in size and completely blocking the cerebral blood vessels; cerebral embolism, in which a blood clot formed by coagulating partially stagnant blood due to an abnormality in blood flow within the heart breaks off and travels through the bloodstream to block relatively small cerebral blood vessels, thereby blocking blood flow; and lacunar cerebral infarction, in which tiny blood vessels in the brain become blocked due to high blood pressure, etc. When acute cerebral infarction occurs, symptoms such as speech impairment, unilateral paralysis (weakness in one side of the body), and visual impairment occur suddenly, and if treatment is delayed, it leaves significant aftereffects; therefore, the main treatment method for acute cerebral infarction aims to remove the blocked blood clot. According to treatment guidelines for acute cerebral infarction, it is known that in the early stages of acute cerebral infarction, treatment with a thrombolytic agent called recombinant tissue plasminogen (tPA) activator can improve the patient's cerebral infarction findings. Thrombolytic agents break down blood clots to allow blood to flow again through blocked blood vessels. However, since the risk of cerebral hemorrhage increases significantly over time with the use of thrombolytic agents, it is generally known that administration is permitted only within 4.5 hours of the onset of cerebral infarction. Therefore, patients whose exact time of onset of cerebral infarction is unknown are excluded from thrombolytic treatment. The American Heart Association and the American Stroke Association (AHA/ASA) recommended that medical professionals use the DWI-FLAIR mismatch [comparison of DWI (Diffusion-Weighted Imaging) positive lesions and FLAIR (Fluid Attenuated Inversion Recovery) negative lesions] to identify cerebral infarction in patients whose time of stroke onset is unknown and who are candidates for thrombolytic therapy. Diffusion-Weighted Imaging (DWI) visualizes the degree of diffusion of water molecules within the human body, and DWI can identify signal changes in areas affected by tissue damage or lesions in the brain based on the movement of water molecules. FLAIR (Fluid Attenuated Inversion Recovery) imaging is a method that suppresses signals associated with water molecules to make other brain structures appear more distinct. When acute cerebral infarction occurs, damaged brain cells exhibit restricted diffusion of water molecules, resulting in dark signals on DWI. On the other hand, for acute cerebral infarction, FLAIR is captured before changes in the affected area are identified, and it has been known that the discrepancy between DWI and FLAIR can be used to identify candidates for thrombolytic therapy. However, related studies have shown low to moderate levels of accuracy, and both imaging techniques require magnetic resonance imaging (MRI) equipment. Due to the nature of MRI equipment, it is not available in all hospitals or emergency rooms, and the relatively long examination time presents a disadvantage that may be unfavorable in emergency situations. Accordingly, there has been a need for technology capable of diagnosing cerebral infarction and predicting the onset