WO2022145541A1 - Procédé et dispositif d'aide au diagnostic de maladie rénale - Google Patents

Procédé et dispositif d'aide au diagnostic de maladie rénale Download PDF

Info

Publication number
WO2022145541A1
WO2022145541A1 PCT/KR2020/019450 KR2020019450W WO2022145541A1 WO 2022145541 A1 WO2022145541 A1 WO 2022145541A1 KR 2020019450 W KR2020019450 W KR 2020019450W WO 2022145541 A1 WO2022145541 A1 WO 2022145541A1
Authority
WO
WIPO (PCT)
Prior art keywords
diagnosis
information
kidney disease
renal disease
fundus image
Prior art date
Application number
PCT/KR2020/019450
Other languages
English (en)
Korean (ko)
Inventor
최태근
이근영
박정탁
Original Assignee
주식회사 메디웨일
연세대학교 산학협력단
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 주식회사 메디웨일, 연세대학교 산학협력단 filed Critical 주식회사 메디웨일
Priority to PCT/KR2020/019450 priority Critical patent/WO2022145541A1/fr
Publication of WO2022145541A1 publication Critical patent/WO2022145541A1/fr

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
    • A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
    • A61B3/12—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/20—Measuring for diagnostic purposes; Identification of persons for measuring urological functions restricted to the evaluation of the urinary system
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment

Definitions

  • the present specification relates to a method and apparatus for assisting in the diagnosis of kidney disease, and to a method and apparatus for assisting in the diagnosis of kidney disease using an artificial neural network model.
  • This specification is derived from research conducted as part of the construction of an open competition platform through the Information and Communication Industry Promotion Agency (NIPA, Korea) created by the Information Promotion Fund of the Ministry of Science and ICT (MSIT, Korea) in 2020. [Task Management No.: A0712-20-1002]
  • Fundus examination can observe abnormalities in the retina, optic nerve, and macula, and is a diagnostic aid frequently used in ophthalmology because the results can be confirmed through relatively simple imaging.
  • the use of the fundus examination is increasing in that it is possible to observe the degree of damage to blood vessels caused by chronic diseases such as high blood pressure and diabetes in a non-invasive way as well as eye diseases.
  • One object of the present invention is to provide a method for assisting diagnosis of kidney disease.
  • Another object of the present invention is to provide a method of assisting in the diagnosis of kidney disease using a neural network model based on a fundus image.
  • a method for assisting in diagnosis of a target kidney disease using a fundus image comprising: acquiring a target fundus image obtained by photographing the fundus of a subject; based on the target fundus image, the fundus image Acquiring the renal disease diagnosis auxiliary information of the subject according to the target fundus image through the renal disease diagnosis auxiliary neural network model for obtaining the diagnosis auxiliary information used for the diagnosis of the target renal disease according to A step of outputting information, wherein the renal disease diagnosis auxiliary information is grade information including a grade selected from among a plurality of grades indicating the risk of the target kidney disease, score information and blood that is numerical information for determining the risk of the target kidney disease
  • a method for assisting in diagnosing kidney disease including at least one of risk information indicating whether a sample corresponds to a target kidney disease risk group may be provided.
  • information that can be used for diagnosing kidney disease may be obtained based on the fundus image.
  • various information that can be used for diagnosing kidney disease may be acquired based on the fundus image.
  • 1 is a view for explaining an example of a method for assisting diagnosis of kidney disease.
  • FIG. 2 is a diagram for describing a diagnosis unit according to an exemplary embodiment.
  • FIG 3 is a view for explaining a method for assisting in diagnosing kidney disease according to an exemplary embodiment.
  • FIG. 4 is a view for explaining a method of assisting in diagnosing kidney disease according to another exemplary embodiment.
  • a method for assisting in diagnosis of a target kidney disease using a fundus image comprising: acquiring a target fundus image obtained by photographing the fundus of a subject; based on the target fundus image, the fundus image Acquiring the renal disease diagnosis auxiliary information of the subject according to the target fundus image through the renal disease diagnosis auxiliary neural network model for obtaining the diagnosis auxiliary information used for the diagnosis of the target renal disease according to A step of outputting information, wherein the renal disease diagnosis auxiliary information is grade information including a grade selected from among a plurality of grades indicating the risk of the target kidney disease, score information and blood that is numerical information for determining the risk of the target kidney disease
  • a method for assisting in diagnosing kidney disease including at least one of risk information indicating whether a sample corresponds to a target kidney disease risk group may be provided.
  • the method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium.
  • the computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination.
  • the program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software.
  • Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks.
  • - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like.
  • Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.
  • the hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.
  • a system, apparatus, method, and the like may be provided for assisting in the diagnosis of kidney disease using a fundus image.
  • a system, apparatus, method, and the like for assisting in the diagnosis of kidney disease by using a neural network model and obtaining diagnostic assistance information useful for the diagnosis of kidney disease based on a fundus image.
  • biomarkers used directly or indirectly for disease diagnosis may be used.
  • a method of managing the risk of a disease may be used in consideration of values of disease-related indexes, scores, or indicators (hereinafter, scores). With respect to diseases diagnosed in consideration of a value such as a score, it may be more efficient to provide a score, not the presence or absence of a disease, because the clinician may directly consider the score to determine the patient's condition or treatment.
  • Kidney disease described herein may refer to a disease related to the kidney.
  • kidney diseases include acute renal failure, acute progressive renal failure, acute nephritis, diabetic renal failure, glomerulonephritis, chronic renal failure, nephrotic syndrome, pyelonephritis, polycystic nephropathy, asymptomatic urinary abnormality, urinary tract infection, renal tubular deficiency, hypertension. It may refer to diseases related to the kidneys, such as kidney disease, nephrolithiasis, and urinary tract obstruction.
  • kidney disease described herein may be accompanied by complications.
  • kidney disease may be accompanied by complications such as uremia, hyperkalemia, volume overload, heart disease, hypertension, anemia, and the like. Kidney disease as described herein may refer to such complications.
  • diagnostic assistance information used for diagnosing a disease may be provided.
  • the diagnostic auxiliary information may include a parameter value related to a kidney disease, a grade indicating a risk of a kidney disease, or information on the presence or absence of a kidney disease.
  • the score to aid in the diagnosis of kidney disease may be a score that can be measured from a subject or a score calculated by combining values measured from the subject and/or personal information of the subject.
  • the score used for the diagnosis of kidney disease may be a score presented by a known kidney disease prediction model.
  • a score that aids in the diagnosis of kidney disease may be a creatinine index indicative of creatinine levels.
  • an auxiliary score for diagnosing kidney disease may be eGFR.
  • the score to aid in the diagnosis of kidney disease may be a secondary score processed based on at least one of creatinine levels or eGFR.
  • diagnostic auxiliary information such as a score to assist in the diagnosis of kidney disease may be used as a criterion for selecting a specific medical treatment or prescription target.
  • the score assisting in the diagnosis of kidney disease may be used to select a target for kidney disease work-up.
  • the score supporting the diagnosis of kidney disease may be used as a prescription criterion for a drug for treating kidney disease.
  • a drug for treating kidney disease may be derived based on a score that aids in the diagnosis of kidney disease.
  • the drug for treating kidney disease may be various drugs for treating kidney disease.
  • the drug for treating kidney disease may include drugs for treating other diseases related to kidney disease (eg, hypertension, diabetes, hyperlipidemia, etc.).
  • the grade that aids in the diagnosis of kidney disease may be at least one grade indicative of a risk of kidney disease. For example, when a score or the like may be used for diagnosis of a disease, a grade may be used in place of or together with a score or the like.
  • the diagnostic assistance information may include a renal disease diagnostic assistance score and/or a renal disease diagnostic assistance grade.
  • the grade may include a normal grade indicating that the subject is normal for the target kidney disease and an abnormal grade indicating that the subject is abnormal for the target kidney disease.
  • the ratings may include a plurality of ratings indicating the degree of risk of the subject for the target kidney disease.
  • Scores and the like described herein may be used for diagnosing a disease.
  • the score and the like may be used to diagnose a patient's current condition and/or to manage the prognosis of a disease.
  • renal disease diagnosis assistance information may be acquired prior to diagnosis of various renal diseases.
  • the acquired renal disease diagnosis auxiliary information may be used for pre-diagnosis to select a subject for a detailed examination prior to a detailed examination for kidney disease.
  • the description is based on the case for the purpose of diagnosing kidney disease, but the content of the invention disclosed in the present specification is not limited thereto.
  • the diagnostic assistance method, apparatus, and system described below use a neural network model to provide a numerical value as diagnostic assistance information related to the disease from a fundus image. Alternatively, it may be similarly applied to all cases of obtaining a grade.
  • a diagnosis assistance system and apparatus for acquiring information based on the presence or absence of a kidney disease or determination based on the fundus image will be described.
  • deep learning to build a neural network model for predicting information related to kidney disease, learning the built model, and predicting information using the learned model, systems and devices that aid in the diagnosis of kidney disease, etc. about it.
  • an auxiliary system for diagnosing kidney disease for acquiring auxiliary information for diagnosis of kidney disease based on a fundus image
  • a renal disease diagnosis assistance system or a device constituting the same may perform diagnostic assistance and/or renal disease diagnosis assistance described throughout this specification.
  • the renal disease diagnosis assistance system may include a learning device, a diagnosis device, and a client device.
  • the learning device learns a neural network model for assisting in diagnosing kidney disease
  • the diagnostic device assisting in diagnosing kidney disease using the learned neural network model
  • the client device acquiring a fundus image, and a kidney generated based on the fundus image
  • the disease diagnosis auxiliary information may be obtained, and the diagnosis auxiliary information may be provided to the user.
  • the renal disease diagnosis assistance system may include a diagnostic device and a client device.
  • the diagnosis apparatus may perform functions of a learning apparatus and/or a server apparatus.
  • the diagnostic device and/or client device may perform the renal disease diagnosis aid described herein.
  • the renal disease diagnosis assistance system may include a mobile device.
  • the mobile device may perform all or some of the operations of the learning device, the diagnostic device, and/or the client device.
  • the mobile device may perform the renal disease diagnosis aid described herein.
  • a diagnostic assistance apparatus may be provided.
  • the diagnostic assistant may include a kidney disease diagnostic assistant.
  • the kidney disease diagnosis assistant may perform a kidney disease diagnosis assistant described herein.
  • the diagnosis assistance device may acquire diagnosis assistance information of a kidney disease based on the fundus image.
  • the diagnostic aid device may be one or more devices including a kidney disease diagnostic aid.
  • the diagnostic assistance device may be a learning device, a diagnostic device, or a client device.
  • the diagnostic assistance device may be a mobile device.
  • the diagnostic auxiliary device may be a server device.
  • the diagnosis assistance device may include a renal disease diagnosis auxiliary unit that assists in the diagnosis of a renal disease described herein.
  • a diagnostic aid device may be included in any device or component described herein.
  • the learning device may assist in diagnosing kidney disease.
  • the learning apparatus may include a kidney disease diagnosis assistant to assist in diagnosing kidney disease.
  • the renal disease diagnosis auxiliary unit of the learning apparatus may train a neural network model for predicting renal disease diagnosis auxiliary information based on the fundus image.
  • the renal disease diagnosis auxiliary unit of the learning apparatus may assist in the diagnosis of renal disease by learning a neural network model for predicting renal disease diagnosis auxiliary information based on the fundus image.
  • the learning unit of the learning device may perform a function of the renal disease diagnosis auxiliary unit assisting in diagnosing kidney disease.
  • the learning unit may include a kidney disease diagnosis assistant.
  • a processor or a learning module of the learning device may perform a function of a kidney disease diagnosis assistant.
  • the processor or the learning module of the learning device may include a kidney disease diagnosis assistant.
  • the controller of the learning apparatus may perform diagnosis assistance for kidney disease described herein.
  • the control unit may include a kidney disease diagnosis assistant.
  • the memory unit (or volatile memory, non-volatile memory, or mass storage device) of the learning device may store a neural network model for assisting in diagnosing kidney disease.
  • the communication unit of the learning apparatus may transmit the learned model or information for driving the learned model to an external device. Alternatively, the learning device may acquire information necessary for learning the neural network model from an external device through the communication unit.
  • the diagnostic device may assist in diagnosing kidney disease.
  • the diagnostic apparatus may include a kidney disease diagnosis assistant to assist in diagnosing kidney disease.
  • the renal disease diagnosis auxiliary unit of the diagnosis apparatus may obtain renal disease diagnosis auxiliary information on the target fundus image by using the learned neural network model.
  • the renal disease diagnosis auxiliary unit of the diagnosis apparatus may obtain the renal disease diagnosis auxiliary information by using a neural network model that outputs the renal disease diagnosis auxiliary information based on the fundus image.
  • the diagnosis unit of the diagnosis apparatus may perform a function of the renal disease diagnosis auxiliary unit that assists in diagnosing kidney disease.
  • the diagnostic unit may include a kidney disease diagnosis assistant.
  • a processor or a diagnosis module of the diagnosis apparatus may perform a function of a kidney disease diagnosis auxiliary unit.
  • a processor or a diagnostic module of the diagnostic apparatus may include a kidney disease diagnosis assistant.
  • the controller of the diagnosis apparatus may assist in diagnosing kidney disease by using the learned neural network model.
  • the control unit of the diagnosis apparatus may perform a function of the renal disease diagnosis auxiliary unit.
  • the control unit of the diagnosis apparatus may include a kidney disease diagnosis auxiliary unit.
  • the memory unit of the diagnosis apparatus may store the learned neural network model to assist in diagnosing kidney disease.
  • the memory unit of the diagnosis apparatus may include a kidney disease diagnosis auxiliary unit.
  • the diagnostic apparatus may communicate with an external device using a communication unit.
  • the diagnosis apparatus may acquire a diagnosis target image from an external device or transmit diagnosis auxiliary information to the external device using the communication unit.
  • the diagnosis apparatus may acquire a neural network model learned from an external device (eg, a learning device) or information necessary for driving the learned neural network model by using the communication unit.
  • the client device may assist in the diagnosis of kidney disease.
  • the client device may include a kidney disease diagnosis assistant to assist in the diagnosis of kidney disease.
  • the renal disease diagnosis assistant of the client device may learn a neural network model, obtain diagnostic assistance information using the neural network model, or provide data (eg, fundus image) necessary for driving the neural network model.
  • the client device may obtain information necessary to assist in the diagnosis of kidney disease from the user, or provide information to assist in diagnosis of kidney disease to the user.
  • the control unit of the client device may include a kidney disease diagnosis assistant.
  • the control unit of the client device may perform a function of a kidney disease diagnosis assistant.
  • the processor of the client device may include a renal disease diagnosis aid or perform renal disease diagnosis aid.
  • the server device may assist in diagnosing kidney disease.
  • the server device may include a kidney disease diagnosis assistant to assist in the diagnosis of kidney disease.
  • the renal disease diagnosis assistant of the server device may store, learn, or drive the neural network model.
  • the server device may store data (eg, fundus image data) necessary for storing, learning, or driving a neural network model that assists in diagnosing kidney disease.
  • the server device may store user information used to assist in diagnosing kidney disease.
  • Kidney disease diagnosis assistance includes learning a renal disease diagnosis auxiliary neural network model using the learning target fundus image, and acquiring renal disease diagnosis auxiliary information based on the diagnosis target fundus image using the learned renal disease diagnosis auxiliary neural network model.
  • the renal disease diagnosis auxiliary neural network model may be a multi-layer neural network model that outputs a renal disease-related diagnosis auxiliary result.
  • the renal disease diagnosis auxiliary neural network model may be a convolutional neural network model that acquires diagnostic auxiliary information based on the fundus image.
  • the renal disease diagnosis auxiliary neural network model may be prepared in the form of an ensemble.
  • the renal disease diagnosis auxiliary neural network model includes a first sub-neural network model that outputs a first result and a second sub-neural network model that outputs a second result, and the obtained diagnostic auxiliary information includes the first result and the second result. It can be decided by considering the results together.
  • Kidney disease diagnosis assistance can be largely divided into learning of a renal disease diagnosis auxiliary neural network model and a diagnosis aid using a learned renal disease diagnosis auxiliary neural network model.
  • Learning the renal disease diagnosis auxiliary neural network model may include acquiring training data and learning the renal disease diagnosis auxiliary neural network model based on the acquired training data.
  • Acquiring the training data may include acquiring fundus image training data.
  • the acquired fundus image learning data may be fundus image learning data labeled with kidney disease diagnosis information. In this regard, it will be described in more detail in the fundus image acquisition section below.
  • Training the renal disease diagnosis auxiliary model may include transforming (or pre-processing) the acquired fundus image in some cases.
  • the renal disease diagnosis auxiliary neural network model may be trained using the transformed fundus image.
  • the fundus image may be converted or pre-processed into a form more suitable for obtaining diagnostic auxiliary information for kidney disease. In this regard, it will be looked at in more detail in the fundus image pre-processing section below.
  • Training the neural network model for diagnosing kidney disease may include predicting a result on unit learning data, comparing the predicted result with a label, and repeating the steps of updating the neural network model several times. In this regard, it will be described in more detail in the training stage of the renal disease diagnosis auxiliary neural network model below.
  • Assisting in the diagnosis of kidney disease by using the neural network model may include acquiring an image of the fundus to be diagnosed and acquiring information of assisting in diagnosis of kidney disease from the fundus image to be diagnosed using the learned neural network model. have.
  • Acquiring the fundus image to be diagnosed may be performed through an imaging unit or a captured image may be acquired through a separate imaging device.
  • Assisting in the diagnosis of kidney disease may include preprocessing the fundus image to be diagnosed.
  • the acquired fundus image to be diagnosed may be pre-processed.
  • Acquiring the diagnosis auxiliary information of the kidney disease may be obtaining the diagnosis auxiliary information on the fundus image preprocessed using the learned neural network model. In this regard, it will be looked at in more detail in the fundus image pre-processing section below.
  • the pre-processing of the fundus image to be diagnosed may include converting or pre-processing the fundus image to be diagnosed into a form more suitable for obtaining diagnostic auxiliary information of kidney disease. In this regard, it will be looked at in more detail in the fundus image pre-processing section below.
  • Acquiring the diagnosis auxiliary information of kidney disease by using the learned neural network model may include acquiring disease presence information, numerical information, grade level, etc. that can be used for diagnosing kidney disease. In this regard, it will be described in more detail in the subsection for diagnosis of kidney disease using a neural network model below.
  • the diagnosis assistance process performs preprocessing on an image (S101), learns a renal disease diagnosis auxiliary neural network model based on the preprocessed image (S103) , a learning process of acquiring the parameters of the learned renal disease diagnosis auxiliary neural network model (S105), and after acquiring the diagnosis target image and preprocessing the diagnosis target image (S401), using the learned renal disease diagnosis auxiliary neural network model (S403)
  • a diagnosis auxiliary process of obtaining renal disease diagnosis auxiliary diagnosis auxiliary information (S405) may be included.
  • the training process of the renal disease diagnosis auxiliary neural network model includes a preprocessing step of preprocessing the fundus image to improve prediction accuracy of renal disease diagnosis information, and learning to train the renal disease diagnosis auxiliary neural network model using the preprocessed fundus image It can include processes.
  • the training process may be performed by the learning device.
  • the diagnosis assistance process using the neural network model assisting in diagnosing kidney disease may include a preprocessing process of preprocessing an input target fundus image and a diagnosis auxiliary process of assisting in diagnosing a kidney disease using the preprocessed fundus image.
  • the diagnostic auxiliary process may be performed by a diagnostic device or a server device.
  • the learned renal disease diagnosis auxiliary neural network model may be used to assist the diagnosis of renal disease.
  • the learned renal disease diagnosis auxiliary neural network model it is possible to assist in the diagnosis of renal disease by acquiring diagnostic auxiliary information useful for the diagnosis of renal disease.
  • the diagnosis device, the client device, the mobile device, or the server device may acquire the diagnosis assistance information based on the fundus image of the patient.
  • Each device includes a diagnosis unit, a control unit, or a processor, and the diagnosis unit, control unit, or processor of each device may acquire auxiliary diagnosis information according to the target fundus image by using the renal disease diagnosis auxiliary neural network model.
  • the diagnosis unit 500 may include a diagnosis request obtaining module 501 , a renal disease diagnosis auxiliary module 503 , and a diagnosis auxiliary information output module 505 .
  • the diagnosis unit 500 and each module shown in FIG. 2 are illustratively described based on a logical configuration, and the diagnosis unit 500 or each module is at least one of various devices or other devices described in this specification. may be included in the device.
  • the diagnosis request obtaining module 501 may obtain a diagnosis request from an external device or a user.
  • the diagnosis request obtaining module 501 may obtain a diagnosis request including a fundus image to be diagnosed.
  • the diagnosis request obtaining module 501 may obtain a diagnosis request including diagnosis auxiliary information identification information for identifying the requested diagnosis auxiliary information.
  • the diagnosis request obtaining module 501 may obtain a diagnosis request and initiate a diagnosis auxiliary information obtaining process.
  • the diagnosis request obtaining module 501 may obtain a fundus image after obtaining a diagnosis request, obtain a diagnosis request including a fundus image, or obtain a diagnosis request after obtaining a fundus image.
  • the renal disease diagnosis auxiliary module 503 may obtain diagnosis auxiliary information by using the learned renal disease diagnosis auxiliary neural network model.
  • the renal disease diagnosis assistance module 503 may acquire diagnosis assistance information when a diagnosis assistance request is obtained.
  • the renal disease diagnosis auxiliary module 503 may acquire a target fundus image and acquire renal disease diagnosis auxiliary information from a neural network model based on the target fundus image.
  • the renal disease diagnosis auxiliary module 503 may obtain the learned neural network module or parameters of the learned neural network module, and use the obtained parameters to obtain diagnostic assistance information according to the target fundus image.
  • the renal disease diagnosis auxiliary module 503 may acquire a target fundus image and acquire disease presence information, grade information, or score information for diagnosing kidney disease.
  • the renal disease diagnosis auxiliary module 503 may further acquire additional information (ie, secondary diagnosis auxiliary information) in addition to the primary renal disease diagnosis auxiliary information directly output from the neural network model.
  • additional information ie, secondary diagnosis auxiliary information
  • the renal disease diagnosis auxiliary module 503 may acquire instruction information or prescription information, which will be described later.
  • the renal disease diagnosis auxiliary module may also acquire a CAM (class activation map) image corresponding to the diagnostic auxiliary information or outputted diagnostic auxiliary information for a disease other than the target disease.
  • the diagnosis auxiliary information output module 505 may obtain diagnosis auxiliary information from the renal disease diagnosis auxiliary module.
  • the diagnostic auxiliary information output module 505 may output diagnostic auxiliary information on kidney disease.
  • the diagnostic auxiliary information obtaining module may transmit the diagnostic auxiliary information to an external device or an external module.
  • the diagnosis auxiliary information may be provided to the user through a user interface or the like.
  • the method for assisting kidney disease diagnosis includes the steps of obtaining a diagnosis request ( S201 ), obtaining diagnostic auxiliary information using a neural network model assisting in diagnosing kidney disease ( S203 ), and It may include outputting the diagnostic auxiliary information (S205).
  • the step ( S201 ) of obtaining the diagnosis auxiliary information by using the renal disease diagnosis auxiliary neural network model may be configured differently depending on the type of the target diagnosis auxiliary information (ie, the diagnosis auxiliary information to be acquired).
  • the renal disease diagnosis auxiliary neural network model used to obtain the diagnosis auxiliary information may be determined according to the type of the target diagnosis auxiliary information.
  • Acquiring the auxiliary diagnosis information using the renal disease diagnosis auxiliary neural network model may include acquiring and processing the diagnosis auxiliary information obtained through the renal disease diagnosis auxiliary neural network model.
  • the obtaining of the auxiliary diagnosis information may further include obtaining the secondary diagnosis auxiliary information obtained based on the primary diagnosis auxiliary information and/or the primary diagnosis auxiliary information directly obtained through the neural network model.
  • the step of outputting the diagnostic auxiliary information may include outputting the diagnostic auxiliary information in a form recognizable by a user.
  • Outputting the diagnostic auxiliary information may include outputting the diagnostic auxiliary information in the form of visual or auditory data.
  • the outputting of the diagnostic auxiliary information may be configured differently depending on the type of the outputted diagnostic auxiliary information. For example, the diagnostic auxiliary information may be output differently according to the type.
  • the method for assisting in diagnosing kidney disease according to an embodiment of the present invention includes acquiring a target fundus image (S310), acquiring information for diagnosing kidney disease of a subject (S330), and blood
  • the step of outputting kidney disease diagnosis auxiliary information of the specimen (S350) may be included.
  • Acquiring the target fundus image ( S310 ) may include acquiring the target fundus image obtained by capturing the fundus of the subject.
  • the step of obtaining the renal disease diagnosis auxiliary information of the subject (S310) is based on the target fundus image, through the renal disease diagnosis auxiliary neural network model for obtaining the diagnosis auxiliary information used for the diagnosis of the target kidney disease according to the fundus image,
  • the method may include acquiring information for diagnosing kidney disease of the subject according to the target fundus image.
  • the renal disease diagnosis auxiliary information includes grade information including a grade selected from among a plurality of grades indicating the risk of the target kidney disease, score information that is numerical information for determining the risk of the target kidney disease, and whether the subject corresponds to the target kidney disease risk group may include at least one of risk information indicating
  • the step of obtaining the renal disease diagnosis auxiliary information of the subject may further include obtaining a class activation map for the renal disease diagnosis auxiliary information according to the target fundus image.
  • the step of outputting the renal disease diagnosis auxiliary information of the subject ( S350 ) may include outputting the obtained renal disease diagnosis auxiliary information.
  • a feature area related to the renal disease diagnosis auxiliary information is generated based on the class activation map for the renal disease diagnosis auxiliary information and provided in a form corresponding to the target fundus image.
  • the method may further include outputting a kidney disease diagnosis auxiliary feature map indicating
  • the step of outputting the renal disease diagnosis auxiliary information of the subject ( S350 ) may include outputting the target fundus image and the renal disease diagnosis auxiliary feature map by overlapping them.
  • the step of outputting the renal disease diagnosis auxiliary information of the subject may further include outputting indication information determined based on the renal disease diagnosis auxiliary information of the subject.
  • the indication information may be determined using a pre-stored renal disease diagnosis auxiliary information-indicative information relationship.
  • the indication information may be determined using a matching table in which indication information according to the renal disease diagnosis auxiliary information is matched.
  • the indication information may be determined according to a predetermined renal disease diagnosis auxiliary information-indicative information relationship, and the renal disease diagnosis auxiliary information-indicative information relation may include possible medical treatment methods for the subject corresponding to the renal disease diagnosis auxiliary information. have.
  • the method for assisting diagnosis of kidney disease may be performed by an apparatus for assisting diagnosis for eye disease.
  • the method of assisting in diagnosing a kidney disease may include assisting in diagnosing an eye disease of a subject according to a target fundus image through a renal disease diagnosis auxiliary neural network model that acquires secondary information for diagnosing eye disease according to the fundus image, based on the target fundus image.
  • the method may further include obtaining information.
  • a renal disease diagnosis auxiliary neural network model that assists in determining whether to prescribe a medical action related to a kidney disease based on the fundus image may be trained.
  • the renal disease diagnosis auxiliary neural network model may be trained as a binary neural network model that classifies a plurality of fundus images into two classes classified according to the necessity of a specific medical action for the subject.
  • the renal disease diagnosis auxiliary neural network model may be trained to classify the fundus image into a first class requiring a specific medical action or a second class not requiring a specific medical action.
  • the renal disease diagnosis adjuvant neural network model provides a fundus image to a first class in which a specific medical action is required in the near future (eg, immediately), a second class in which a specific medical operation is required within a predetermined period (eg, within 3 years), a specific It can be learned to classify into a third class that does not require medical care.
  • the specific medical practice may be a medical treatment or prescription related to the aforementioned kidney disease.
  • a specific medical practice may include a drug or non-pharmacological treatment recommended for improvement of a target disease in a subject.
  • a specific medical practice may be the administration of a specific drug or agent or a prescription thereof.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Biomedical Technology (AREA)
  • Medical Informatics (AREA)
  • Physics & Mathematics (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Pathology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Veterinary Medicine (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Physiology (AREA)
  • Radiology & Medical Imaging (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Urology & Nephrology (AREA)
  • Ophthalmology & Optometry (AREA)
  • Eye Examination Apparatus (AREA)

Abstract

L'invention concerne un procédé et un dispositif d'aide au diagnostic de maladie rénale. Un aspect de la présente invention concerne un procédé d'aide au diagnostic d'une maladie rénale chez un sujet à l'aide d'une image de fond d'oeil, le procédé comprenant les étapes consistant à : obtenir une image de fond d'oeil de sujet qui est une image capturée du fond d'oeil d'un sujet ; obtenir des informations d'aide au diagnostic de maladie rénale du sujet conformément à l'image de fond de l'oeil du sujet, au moyen d'un modèle de réseau de neurones artificiels d'aide au diagnostic de maladie rénale obtenant des informations d'aide au diagnostic destinées à être utilisées dans le diagnostic de la maladie rénale du sujet conformément à l'image de fond d'oeil, sur la base de l'image de fond d'oeil du sujet ; et délivrer en sortie les informations d'aide au diagnostic de maladie rénale du sujet, les informations d'aide au diagnostic de maladie rénale constituant au moins l'une des informations suivantes : des informations de niveau incluant un niveau sélectionné parmi une pluralité de niveaux indiquant le risque de maladie rénale du sujet ; des informations de score qui sont des informations numériques pour déterminer le score de risque de maladie rénale du sujet ; et des informations de risque indiquant si le sujet appartient ou non à un groupe de risques de maladie rénale.
PCT/KR2020/019450 2020-12-30 2020-12-30 Procédé et dispositif d'aide au diagnostic de maladie rénale WO2022145541A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/KR2020/019450 WO2022145541A1 (fr) 2020-12-30 2020-12-30 Procédé et dispositif d'aide au diagnostic de maladie rénale

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/KR2020/019450 WO2022145541A1 (fr) 2020-12-30 2020-12-30 Procédé et dispositif d'aide au diagnostic de maladie rénale

Publications (1)

Publication Number Publication Date
WO2022145541A1 true WO2022145541A1 (fr) 2022-07-07

Family

ID=82259463

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2020/019450 WO2022145541A1 (fr) 2020-12-30 2020-12-30 Procédé et dispositif d'aide au diagnostic de maladie rénale

Country Status (1)

Country Link
WO (1) WO2022145541A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024146910A1 (fr) * 2023-01-05 2024-07-11 Carl Zeiss Meditec, Inc. Procédé et système de stadification d'une maladie rénale diabétique par apprentissage profond

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150254524A1 (en) * 2012-04-11 2015-09-10 University Of Flordia Research Foundation, Inc. System and method for analyzing random patterns
CN107203778A (zh) * 2017-05-05 2017-09-26 平安科技（深圳）有限公司 视网膜病变程度等级检测系统及方法
KR20200005411A (ko) * 2018-07-06 2020-01-15 연세대학교 산학협력단 심혈관 질병 진단 보조 방법 및 장치
CN111435612A (zh) * 2018-12-26 2020-07-21 福州依影健康科技有限公司 一种移动医疗个性化健康服务的方法和系统
WO2020200087A1 (fr) * 2019-03-29 2020-10-08 Ai Technologies Inc. Détection basée sur une image de maladies ophtalmiques et systémiques

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150254524A1 (en) * 2012-04-11 2015-09-10 University Of Flordia Research Foundation, Inc. System and method for analyzing random patterns
CN107203778A (zh) * 2017-05-05 2017-09-26 平安科技（深圳）有限公司 视网膜病变程度等级检测系统及方法
KR20200005411A (ko) * 2018-07-06 2020-01-15 연세대학교 산학협력단 심혈관 질병 진단 보조 방법 및 장치
CN111435612A (zh) * 2018-12-26 2020-07-21 福州依影健康科技有限公司 一种移动医疗个性化健康服务的方法和系统
WO2020200087A1 (fr) * 2019-03-29 2020-10-08 Ai Technologies Inc. Détection basée sur une image de maladies ophtalmiques et systémiques

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024146910A1 (fr) * 2023-01-05 2024-07-11 Carl Zeiss Meditec, Inc. Procédé et système de stadification d'une maladie rénale diabétique par apprentissage profond

Similar Documents

Publication Publication Date Title
JP4416828B2 (ja) 神経精神病の診断を生成するためのコンピュータ実装方法、コンピュータ・プログラム製品、装置及びシステム
WO2018106005A1 (fr) Système de diagnostic d'une maladie à l'aide d'un réseau neuronal et procédé associé
WO2019031794A1 (fr) Procédé permettant de générer un résultat de prédiction pour prédire une occurrence de symptômes fatals d'un sujet à l'avance et dispositif utilisant ce dernier
KR20190066689A (ko) 진료 보조 시스템 및 방법
WO2020111754A9 (fr) Procédé pour fournir un système de diagnostic utilisant l'apprentissage semi-supervisé, et système de diagnostic l'utilisant
WO2022145541A1 (fr) Procédé et dispositif d'aide au diagnostic de maladie rénale
WO2021040327A1 (fr) Appareil et procédé de prédiction du facteur de risque cardiovasculaire
WO2022158843A1 (fr) Procédé d'affinage d'image d'échantillon de tissu, et système informatique le mettant en œuvre
WO2020213826A1 (fr) Procédé auxiliaire permettant le diagnostic de symptômes des voies urinaires inférieures
Naik et al. Efficient diabetic retinopathy detection using convolutional neural network and data augmentation
WO2022114791A1 (fr) Système de prédiction de risque de maladie basé sur des mégadonnées, procédé de prédiction et programme
WO2022010282A1 (fr) Procédé et dispositif basés sur une caractéristique vocale pour une prédiction de la maladie d'alzheimer
WO2021201582A1 (fr) Procédé et dispositif permettant d'analyser des causes d'une lésion cutanée
KR20200002201A (ko) 인공지능을 통한 스마트 헬스케어 시스템
WO2020246676A1 (fr) Système de diagnostic automatique du cancer du col de l'utérus
KR20220104582A (ko) 심전도를 기반으로 하는 패혈증 진단 시스템
WO2023121051A1 (fr) Procédé de fourniture d'informations de patient, appareil de fourniture d'informations de patient et support d'enregistrement lisible par ordinateur
CN113053514A (zh) 基于5g通信技术的智慧城市医联体系统
CN117011892A (zh) 一种基于深度学习的超声图像自动分析与诊断系统
WO2017010612A1 (fr) Système et méthode de prédiction de diagnostic pathologique reposant sur une analyse d'image médicale
WO2019164273A1 (fr) Méthode et dispositif de prédiction de temps de chirurgie sur la base d'une image chirurgicale
WO2022145542A1 (fr) Procédé et dispositif d'aide au diagnostic d'une maladie cardiovasculaire
WO2022103140A1 (fr) Dispositif d'interprétation d'images médicales et procédé d'interprétation d'images médicales
EP3945527A1 (fr) Vérification automatique de résultat de test
WO2022065925A1 (fr) Procédé et appareil pour aider au diagnostic d'un avc d'embolie cardiaque par l'utilisation d'images radiographiques de la poitrine

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20968087

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20968087

Country of ref document: EP

Kind code of ref document: A1