WO2022139321A1 - Système de rapport d'image de pathologie du cancer de la prostate à base d'intelligence artificielle - Google Patents

Système de rapport d'image de pathologie du cancer de la prostate à base d'intelligence artificielle Download PDF

Info

Publication number
WO2022139321A1
WO2022139321A1 PCT/KR2021/019168 KR2021019168W WO2022139321A1 WO 2022139321 A1 WO2022139321 A1 WO 2022139321A1 KR 2021019168 W KR2021019168 W KR 2021019168W WO 2022139321 A1 WO2022139321 A1 WO 2022139321A1
Authority
WO
WIPO (PCT)
Prior art keywords
prostate
image data
image
gleason
report
Prior art date
Application number
PCT/KR2021/019168
Other languages
English (en)
Korean (ko)
Inventor
이명재
강신욱
김원태
김동민
김동석
Original Assignee
(주)제이엘케이
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
Priority claimed from KR1020200183795A external-priority patent/KR102684355B1/ko
Application filed by (주)제이엘케이 filed Critical (주)제이엘케이
Publication of WO2022139321A1 publication Critical patent/WO2022139321A1/fr

Links

Images

Classifications

    • 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
    • G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
    • 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
    • A61B5/0033—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room
    • A61B5/004—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room adapted for image acquisition of a particular organ or body part
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves 
    • A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves  involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/43—Detecting, measuring or recording for evaluating the reproductive systems
    • A61B5/4375—Detecting, measuring or recording for evaluating the reproductive systems for evaluating the male reproductive system
    • A61B5/4381—Prostate evaluation or disorder diagnosis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
    • A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • 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
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • 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
    • 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
    • G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients

Definitions

  • the present invention relates to a system for calculating a prostate cancer grade from individual prostate histopathology images and outputting a prostate cancer diagnosis result by merging individual prostate histopathological images.
  • Prostate is a male genital gland, which produces prostatic fluid and discharges it to the outside of the body.
  • Prostate fluid is a slightly alkaline milky white liquid that kills bacteria in the urinary tract and neutralizes acid in a woman's vagina to help sperm survive.
  • Prostate cancer is a malignant tumor that starts from the periphery of the prostate, and as the cancer progresses, it can compress the urethra or cause other urological problems.
  • metastasis to central parts of the body, such as the spine or pelvic bones, can occur, causing serious complications.
  • prostate cancer Although the exact cause of prostate cancer is not known, it is rapidly increasing as an aging society progresses and lifestyles become westernized. In particular, although the incidence of prostate cancer is uncommon before the age of 50, the incidence rate rapidly increases after the age of 50.
  • Prostate cancer can be diagnosed through a transrectal resin test, a serum prostate-specific antigen test, a transrectal ultrasound test, and a prostate biopsy.
  • Double prostate biopsy is to classify prostate cancer by observing prostate histopathological images with the naked eye after a prostate biopsy in which a tissue of the prostate is removed and subjected to a microscopic examination.
  • the present invention receives a first image that is an image displayed by the prostate tissue image data, calculates a probability and a Gleason score for each first to fifth Gleason grade, and a prostate tissue classification unit for generating a first report including probabilities and Gleason scores for each of the first to fifth Gleason grades; a biopsy region designator for matching the prostate tissue image data on the entire prostate image data; For all the prostate tissue image data, the sum and percentage of the probabilities for each of the first to fifth Gleason grades are calculated, respectively, and the calculated percentage of the probabilities for each of the first to fifth Gleason grades is calculated as the second of the entire prostate image data.
  • a prostate cancer pathology imaging report system including a pathology imaging unit designated by probability for each grade of 1 to 5 Gleason.
  • the prostate tissue image data is a sample obtained by taking a portion of the prostate tissue under a microscope, and then processing it in a two-dimensional image file format, and the entire prostate image data is a computed tomography image of the prostate ( CT), positron emission tomography imaging (PET), single photon tomography imaging (SPECT), magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), It provides a prostate cancer pathology image report system that is processed into a dimensional image file format.
  • CT computed tomography image of the prostate
  • PET positron emission tomography imaging
  • SPECT single photon tomography imaging
  • MRI magnetic resonance imaging
  • fMRI functional magnetic resonance imaging
  • the prostate tissue classification unit includes a convolutional neural network, wherein the convolutional neural network includes an input layer in the form of a two-dimensional matrix, 1st to Nth filters, 1st to Nth convolutional layers, A prostate cancer pathology imaging report system, comprising 1st to Nth pooling layers, 1st to Mth fully connected layers each in a one-dimensional vector form, and 1st to Nth bias, each of which is one value with an output layer. to provide.
  • the prostate tissue classifier includes: Provided is a prostate cancer pathology imaging report system, in which pixel values of an image are input to the corresponding elements of the input layer.
  • the output layer provides a prostate cancer pathology image report system, including five elements representing probabilities for each of the first to fifth Gleason grades of the prostate tissue image data, respectively.
  • prostate tissue classifier by adding the element of the output layer having the largest value and the element of the output layer having the second largest value to calculate the Gleason score of the prostate tissue image data, prostate cancer pathology Provides a video report system.
  • a prostate cancer pathology imaging report system that receives an input for positioning in
  • the biopsy region designation unit acquires coordinates of a location where the first image is located on the second image to generate biopsy region information of the prostate with respect to the prostate tissue image data, or the first report To provide a prostate cancer pathology image report system, which acquires coordinates of a location on an image and generates information on a biopsy area of the prostate with respect to the prostate tissue image data.
  • the pathology imaging unit is configured to add a Gleason grade having a largest value and a Gleason grade having a second largest value among probabilities for each of the first to fifth Gleason grades of the entire prostate image data to obtain the entire prostate image data
  • a prostate cancer pathology imaging report system which calculates a Gleason score of , and generates a second report including the first to fifth Gleason grade probabilities of the entire prostate image data and the Gleason score.
  • the report output unit provides a prostate cancer pathology image report system that displays the first report and the second report for each of the prostate tissue image data.
  • a prostate cancer grade may be calculated from individual prostate histopathology images, and results of diagnosing prostate cancer may be output by merging the individual prostate histopathological images.
  • the user can conveniently designate a biopsy area of an individual prostate histopathology image by dragging an image or a report of an individual prostate histopathology image and locating it on the entire prostate image.
  • FIG. 1 is a block diagram schematically illustrating a prostate cancer pathology image report system according to an embodiment of the present invention.
  • FIG. 2 is a block diagram schematically illustrating a convolutional neural network according to an embodiment of the present invention.
  • FIG. 3 and 4 are schematic views of a first image and a first report of prostate tissue image data, respectively, according to an embodiment of the present invention.
  • FIG. 5 is a diagram schematically illustrating a second image of whole prostate image data and a first report of individual prostate tissue image data according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a second image and a second report of whole prostate image data according to an embodiment of the present invention.
  • FIG. 1 is a block diagram schematically illustrating a prostate cancer pathology image report system according to an embodiment of the present invention.
  • Prostate cancer pathology image report system 100 is not limited to any computing device capable of reading and processing binary data, such as a server computer, a desktop computer, and a smart phone. can
  • a prostate cancer pathology image report system 100 includes a pathological image input unit 110 , a storage unit 120 , a prostate tissue classification unit 130 , a user input unit 140 , It may include a report output unit 150 , a biopsy region designation unit 160 , and a pathological imaging unit 170 .
  • the pathology image input unit 110 may receive the prostate tissue image data TI and the whole prostate image data PI from the outside of the prostate cancer pathology image report system 100 .
  • the prostate tissue image data (TI) may be obtained by photographing a sample obtained by collecting a portion of prostate tissue under a microscope, and then processing it into a two-dimensional image file format.
  • Prostate whole imaging data is computed tomography imaging (CT), positron emission tomography imaging (PET), single photon tomography imaging (SPECT), magnetic resonance imaging (MRI), and functional magnetic resonance imaging of the prostate.
  • CT computed tomography imaging
  • PET positron emission tomography imaging
  • SPECT single photon tomography imaging
  • MRI magnetic resonance imaging
  • fMRI ultrasound image, after taking any one medical image, may be processed into a two-dimensional image file format.
  • prostate tissue image data and whole prostate image data (PI) are respectively BMP (bitmap), JPG (Joint Photographic Experts Group), PNG (Portable Network Graphics), GIF (Graphics Interchange Format), TIFF (TIFF). Tagged Image File Format) may be processed into any one image file format.
  • BMP bitmap
  • JPG Joint Photographic Experts Group
  • PNG Portable Network Graphics
  • GIF Graphics Interchange Format
  • TIFF TIFF
  • Tagged Image File Format may be processed into any one image file format.
  • the prostate tissue image data TI received by the pathological image input unit 110 may display a two-dimensional image in which a horizontal size and a vertical size are constant.
  • the prostate tissue image data TI received by the pathology image input unit 110 may display a two-dimensional image having a horizontal size of 256 pixels and a vertical size of 256 pixels.
  • Each of the prostate tissue image data TI and the whole prostate image data PI may include metadata about a subject for prostate biopsy.
  • the metadata may include information such as a name, date of birth, and gender of a subject for prostate biopsy.
  • the pathology image input unit 110 includes the prostate cancer pathology image report system 100 and a personal area network (PAN), a local area network (LAN), a metropolitan area network (MAN), and a wide area network.
  • PAN personal area network
  • LAN local area network
  • MAN metropolitan area network
  • Protocols such as Transmission Control Protocol/Internet Protocol (TCP/IP), Server Message Block (SMB), Common Internet File System (CIFS), and Network File System (NFS) from other computing devices connected by a Wide Area Network (WAN).
  • TCP/IP Transmission Control Protocol/Internet Protocol
  • SMB Server Message Block
  • CIFS Common Internet File System
  • NFS Network File System
  • the prostate tissue image data TI and the whole prostate image data PI may be received using other communication protocols.
  • the pathological image input unit 110 includes a serial port, a parallel port, a Small Computer System Interface (SCSI), a Universal Serial Bus (USB), an IEEE 1394, an Advanced Technology Attachment (ATA), and a Serial (SATA). Advanced Technology Attachment), M.2, PCI (Peripheral Component Interconnect Bus), PCI-Express, etc. or other data input/output terminals connected to the prostate tissue image data (TI) and whole prostate image data (PI) can receive
  • the storage unit 120 may store prostate tissue image data TI and whole prostate image data PI.
  • all data stored in the storage unit 120 may be loaded and used in the remaining components except for the storage unit 120 .
  • the storage unit 120 may store all data generated by the components other than the storage unit 120 .
  • the storage unit 120 may include a storage device to store data.
  • Storage devices include hard disk drives, optical disc drives, magnetic tapes, floppy disks, flash memory, solid state drives (SSDs), and the like. It may be a non-volatile memory device, a volatile memory device such as a random access memory (RAM), or another type of memory device.
  • RAM random access memory
  • the prostate tissue classifier 130 may calculate a probability and a Gleason score for each Gleason grade from the prostate tissue image data TI.
  • the Gleason pattern may indicate the degree of differentiation of prostate cancer.
  • Gleason's grade is grade 1 with "densely clustered, single, separate, rounded, monomorphic lines: well-distinguished borders of tumors" and "single, isolated, rounded, Relatively monomorphic glands separated in stromal layers up to the size of a single gland: Grade 2 Gleason grade 2 with relatively distinct tumor borders, and “single, isolated, irregular lines of multiple sizes” : A somatic or papillary tumor with indistinct borders, grade 3 Gleason, and “a tumor with fused glands with invasive cords, papillary, somatic or solid small tumors” Composed of glands: cells of grade 4 Gleason, characterized as small, black or translucent, and composed of cords or plaques of tumor cells invading the stromal layer with few glands on a comedonal tumor background It can be divided into a fifth Gleason grade (grade 5).
  • the Gleason Score may be calculated by adding up the Gleason grade (GP), which appears most frequently in prostate tissue, and the Gleason grade (GP), which appears most frequently in the next.
  • the prostate tissue classifier 130 may include a machine learning model to receive the prostate tissue image data TI and to calculate probabilities for each of the first to fifth Gleason grades (GP).
  • the prostate tissue classifier 130 may include a convolutional neural network 131 .
  • FIG. 2 is a block diagram schematically illustrating a convolutional neural network 131 according to an embodiment of the present invention.
  • the convolutional neural network 131 includes an input layer (IL), first to N-th filters (F 1 to F N ), first to N-th biases (B 1 to B N ), and first to N-th filters.
  • Convolutional layers (CL 1 to CL N ), first to Nth pooling layers (PL 1 to PL N ), first to Mth fully connected layers (FCL 1 to FCL M ), and an output layer (OL) may include
  • the input layer IL may be in the form of a two-dimensional matrix.
  • the number of rows of the input layer IL may be the same as a vertical size of the first image I1 that is an image displayed by the prostate tissue image data TI, and the number of columns may be the same as a horizontal size of the image.
  • the prostate tissue classifier 130 may input the pixel values of the first image I1 into elements of the corresponding input layer IL, respectively.
  • Each of the first to Nth filters F 1 to F N may be in the form of a two-dimensional matrix.
  • Each of the first to Nth biases B 1 to B N may be one value.
  • the prostate tissue classifier 130 performs a convolution operation between the input layer (IL) and the first filter (F 1 ), and then adds a first bias (B 1 ) to each element in a two-dimensional matrix form.
  • a first convolutional layer CL 1 may be generated.
  • the number of rows and columns of the first convolutional layer CL 1 may be the same as the number of rows and columns of the input layer IL, respectively.
  • the prostate tissue classifier 130 divides the first convolutional layer (CL 1 ) into a region having two rows and two columns, and then selects an element having a maximum value in the region to form a two-dimensional matrix.
  • a first pooling layer PL 1 may be generated.
  • the number of rows and columns of the first pooling layer PL 1 may be 1/2 of the number of rows and columns of the first convolutional layer CL 1 , respectively.
  • the prostate tissue classifier 130 performs a convolution operation between the i-1 th pooling layer (PL i-1 ) and the i th filter (Fi ), and then applies the i th bias (B i ) to each element.
  • the i-th convolution layer (CL i ) in the form of a two-dimensional matrix may be generated.
  • the number of rows and columns of the i-th convolutional layer CL i may be the same as the number of rows and columns of the i-1 th pooling layer PL i-1 , respectively. (In this case, 2 ⁇ i ⁇ N, i is a natural number.)
  • the prostate tissue classifier 130 divides the i-th convolutional layer (CL i ) into a region having two rows and two columns, and then selects an element having a maximum value in the region to form a two-dimensional matrix.
  • An i-th pooling layer PL i may be generated.
  • the number of rows and columns of the i-th pooling layer PL i may be 1/2 of the number of rows and columns of the i-th convolutional layer CL i , respectively. (In this case, 2 ⁇ i ⁇ N, i is a natural number.)
  • the prostate tissue classifier 130 may generate the N-th pooling layer PL N by sequentially generating the convolutional layer CL i and the pooling layer PL i .
  • Each of the first to Mth fully connected layers FCL 1 to FCL M may be in the form of a one-dimensional vector.
  • the prostate tissue classifier 130 may input elements of the first fully connected layer FCL 1 corresponding to the elements of the Nth pooling layer PL N , respectively.
  • the prostate tissue classifying unit 130 is a value obtained by multiplying an element of a j-th fully connected layer (FCL j ) by an element of a corresponding j-1th fully connected layer (FCL j-1 ) and a weight therebetween; After adding all the bias values of the elements of the j-th fully connected layer FCL j , the elements of the j-th fully connected layer FCL j may be calculated by inputting them into an activation function. (In this case, 2 ⁇ j ⁇ M, j is a natural number.)
  • the prostate tissue classifier 130 may generate the M-th fully connected layer (FCL M ) by sequentially generating the fully connected layer (FCL j ).
  • the prostate tissue classifying unit 130 is a value obtained by multiplying an element of a j-th fully connected layer (FCL j ) by an element of a corresponding j-1th fully connected layer (FCL j-1 ) and a weight therebetween; After adding all the bias values of the elements of the j-th fully connected layer FCL j , the elements of the j-th fully connected layer FCL j may be calculated by inputting them into an activation function. (In this case, 2 ⁇ j ⁇ M, j is a natural number.)
  • the prostate tissue classifier 130 may generate the M-th fully connected layer (FCL M ) by sequentially generating the fully connected layer (FCL j ).
  • the output layer OL may be in the form of a one-dimensional vector.
  • the output layer OL may include five elements, and the five elements may represent probabilities of becoming first to fifth Gleason grades (grade 1 to grade 5), respectively.
  • Prostate tissue classification unit 130 for the element of the output layer (OL), the value obtained by multiplying the element of the corresponding M-th fully connected layer (FCL M ) and a weight therebetween, and the element of the output layer (OL) After adding up all the bias values, we can input them into the activation function to calculate the elements of the output layer (OL).
  • the prostate tissue classifier 130 is configured to add the element of the output layer OL having the largest value and the element of the output layer OL having the second largest value to add the input Gleason of the prostate tissue image data TI.
  • a score (GS) can be calculated.
  • the prostate tissue classification unit 130 the first to fifth Gleason grade (grade 1 to grade 5) probability according to the prostate tissue image data (TI), the Gleason score (GS), and metadata about the subject for prostate biopsy
  • the first report R1 may be generated by including one or more of them.
  • the user input unit 140 may be an input device capable of receiving an input from a user, for example, an input device such as a mouse, a touchpad, and a joystick.
  • the user input unit 140 displays the first image I1, which is an image displayed by the individual prostate tissue image data TI, from the user, and the second image I2, which is an image displayed by the entire prostate image data PI, from the user.
  • a corresponding input (DRAG) can be received.
  • the user input unit 140 may include a first report (R1) including first to fifth Gleason grade (grade 1 to grade 5) probability and Gleason score (GS) according to individual prostate tissue image data (TI) from the user. , may receive an input DRAG corresponding to the second image I2 of the entire prostate image data PI.
  • R1 first to fifth Gleason grade (grade 1 to grade 5) probability and Gleason score (GS) according to individual prostate tissue image data (TI) from the user.
  • TI prostate tissue image data
  • the report output unit 150 may be an output device capable of displaying an image and a report to the user, for example, a display device.
  • the report output unit 150 includes a first image I1 of individual prostate tissue image data TI, a second image I2 of whole prostate image data PI, and individual prostate tissue image data TI.
  • One of the first report (R1), the probabilities of grade 1 to grade 5 of the whole prostate imaging data (PI), the Gleason score (GS), and metadata about the subject for prostate biopsy The second report R2 including the above may be displayed.
  • the biopsy region designator 160 may correspond to the prostate tissue image data TI on the entire prostate image data PI.
  • the user may drag the first image I1 of the individual prostate tissue image data TI through the user input unit 140 to be positioned on the second image I2 of the entire prostate tissue image data PI. have.
  • the biopsy region designator 160 acquires coordinates of a location where the first image I1 of the individual prostate tissue image data TI is located on the second image I2 of the entire prostate image data PI, Biopsy region information of the prostate with respect to the prostate tissue image data TI may be generated.
  • the user may drag the first report R1 of the individual prostate tissue image data TI through the user input unit 140 to be positioned on the second image I2 of the entire prostate tissue image data PI.
  • the biopsy region designator 160 obtains coordinates of a location where the first report R1 of the individual prostate tissue image data TI is located on the second image I2 of the whole prostate image data PI, Biopsy region information of the prostate with respect to the prostate tissue image data TI may be generated.
  • the pathology imaging unit 170 may calculate a second report R2 of the entire prostate image data PI from the first report R1 of the individual prostate tissue image data TI.
  • the pathology imaging unit 170 may calculate the sum and percentage of probabilities for each of the first to fifth Gleason grades (grades 1 to 5) for all individual prostate tissue image data TI. For example, the probability of the first Gleason grade (grade 1) of all individual prostate tissue imaging data (TI) is added up, and then divided by “number of prostate tissue imaging data (TI) x 100” to calculate the percentage of the first Gleason grade can do.
  • the user input unit 140 receives the first image I1 or the first report R1 of the individual prostate tissue image data TI located on the second image I2 of the whole prostate image data PI from the user. It can take input to select more than one.
  • the pathological imaging unit 170 performs first to fifth Gleason grades with respect to one or more individual prostate tissue image data TI including the first image I1 or the first report R1 selected by the user.
  • the total and percentage of probabilities can be calculated for each grade 1 to 5).
  • the pathology imaging unit 170 calculates the percentage of the probabilities for each of the first to fifth Gleason grades (grade 1 to grade 5), and sets the first to fifth Gleason grades (grade 1 to grade 5) of the entire prostate image data (PI). ) can be specified as a probability.
  • the pathology imaging unit 170 may include a Gleason grade (GP) having a largest value among first to fifth Gleason grade probabilities of the entire prostate image data (PI), a Gleason grade (GP) having the largest value, and a second highest A Gleason score (GS) of the whole prostate image data (PI) may be calculated by adding a Gleason grade (GP) having a value.
  • GP Gleason grade
  • GS Gleason score
  • the pathological imaging unit 170 is configured to include among the first to fifth Gleason grade probabilities of the entire prostate image data (PI), the Gleason score (GS), and metadata about the subject for prostate biopsy.
  • the second report R2 may be generated including one or more.
  • FIG 3 and 4 are schematic views of a first image I1 and a first report R1 of prostate tissue image data TI, respectively, according to an embodiment of the present invention.
  • FIG. 3 is a first image (I1) of prostate tissue image data (TI), and after photographing a sample obtained by collecting a portion of prostate tissue from a subject for prostate biopsy under a microscope, it is processed and displayed in a two-dimensional image file format. can see.
  • R1 prostate tissue image data
  • TI prostate tissue image data
  • Normal Gland the probability of 1st Gleason grade
  • grade 2nd Gleason grade grade 2nd Gleason grade
  • 3rd to 5th Gleason grades You can see that the probabilities (grade 3 ⁇ grade 5) are displayed respectively.
  • FIG. 5 is a diagram schematically illustrating a second image I2 of whole prostate image data PI and a first report R1 of individual prostate tissue image data TI, according to an embodiment of the present invention.
  • 6 is a schematic diagram of a second image I2 and a second report R2 of the entire prostate image data PI according to an embodiment of the present invention.
  • FIG. 5 it can be seen that the second image I2 of the entire prostate image data (PI) is displayed in the left region, which is processed and displayed in a two-dimensional image file format after imaging the prostate from a subject for prostate biopsy. did it
  • a second report R2 of the entire prostate image data PI may be displayed in the middle region.
  • the user drags the first report R1 of the individual prostate tissue image data TI located in the right region through the user input unit 140, respectively, to obtain a second image ( It can be positioned on I2).
  • the Gleason score GS of the individual prostate tissue image data TI may be displayed on the second image I2 of the whole prostate image data PI.
  • the second report (R2) of the entire prostate image data (PI) is the Normal Gland, which is the sum of the probabilities of the 1st Gleason grade (grade 1) and the 2nd Gleason grade (grade 2), and the 3rd to 5th Gleason grades. You can see that the probabilities (grade 3 ⁇ grade 5) are displayed respectively.

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Public Health (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Physics & Mathematics (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Radiology & Medical Imaging (AREA)
  • Data Mining & Analysis (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Optics & Photonics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Reproductive Health (AREA)
  • Gynecology & Obstetrics (AREA)
  • Computational Linguistics (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

La présente invention se rapporte à un système qui permet de calculer un grade du cancer de la prostate à partir d'images individuelles d'histopathologie de la prostate, et de sortir un résultat de diagnostic du cancer de la prostate par fusion des images individuelles d'histopathologie de la prostate. Le système de rapport d'image de pathologie du cancer de la prostate de la présente invention comprend : une unité de classification de tissu de la prostate pour calculer la probabilité et le score de Gleason pour chaque grade de Gleason à partir des données d'image de tissu de la prostate et pour générer un premier rapport ; une unité de désignation de région de biopsie pour mettre en correspondance les données d'image de tissu de la prostate avec toutes les données d'image de la prostate ; une unité de diagnostic d'image pathologique pour calculer la probabilité et le score de Gleason pour chaque grade de Gleason de toutes les données d'image de la prostate.
PCT/KR2021/019168 2020-12-24 2021-12-16 Système de rapport d'image de pathologie du cancer de la prostate à base d'intelligence artificielle WO2022139321A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR1020200183795A KR102684355B1 (ko) 2020-12-24 인공지능 기반의 전립선암 병리 영상 레포트 시스템
KR10-2020-0183795 2020-12-24

Publications (1)

Publication Number Publication Date
WO2022139321A1 true WO2022139321A1 (fr) 2022-06-30

Family

ID=82159644

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2021/019168 WO2022139321A1 (fr) 2020-12-24 2021-12-16 Système de rapport d'image de pathologie du cancer de la prostate à base d'intelligence artificielle

Country Status (1)

Country Link
WO (1) WO2022139321A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090161928A1 (en) * 2007-12-06 2009-06-25 Siemens Corporate Research, Inc. System and method for unsupervised detection and gleason grading of prostate cancer whole mounts using nir fluorscence
JP2019148473A (ja) * 2018-02-27 2019-09-05 シスメックス株式会社 画像解析方法、画像解析装置、プログラム、学習済み深層学習アルゴリズムの製造方法および学習済み深層学習アルゴリズム
KR20190106403A (ko) * 2018-03-09 2019-09-18 연세대학교 산학협력단 질환 예측 방법 및 이를 이용한 질환 예측 디바이스
JP6611612B2 (ja) * 2013-03-15 2019-11-27 シナプティヴ メディカル （バルバドス） インコーポレイテッド 外科データのイントラモーダル同期化
KR20200016666A (ko) * 2018-08-07 2020-02-17 주식회사 딥바이오 진단 결과 생성 시스템 및 방법

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090161928A1 (en) * 2007-12-06 2009-06-25 Siemens Corporate Research, Inc. System and method for unsupervised detection and gleason grading of prostate cancer whole mounts using nir fluorscence
JP6611612B2 (ja) * 2013-03-15 2019-11-27 シナプティヴ メディカル （バルバドス） インコーポレイテッド 外科データのイントラモーダル同期化
JP2019148473A (ja) * 2018-02-27 2019-09-05 シスメックス株式会社 画像解析方法、画像解析装置、プログラム、学習済み深層学習アルゴリズムの製造方法および学習済み深層学習アルゴリズム
KR20190106403A (ko) * 2018-03-09 2019-09-18 연세대학교 산학협력단 질환 예측 방법 및 이를 이용한 질환 예측 디바이스
KR20200016666A (ko) * 2018-08-07 2020-02-17 주식회사 딥바이오 진단 결과 생성 시스템 및 방법

Also Published As

Publication number Publication date
KR20220093017A (ko) 2022-07-05

Similar Documents

Publication Publication Date Title
Gatos et al. Temporal stability assessment in shear wave elasticity images validated by deep learning neural network for chronic liver disease fibrosis stage assessment
RU2543563C2 (ru) Системы и способы поддержки клинических решений
US20120283574A1 (en) Diagnosis Support System Providing Guidance to a User by Automated Retrieval of Similar Cancer Images with User Feedback
Chauhan et al. An overview of image processing for dental diagnosis
Dong et al. One step further into the blackbox: a pilot study of how to build more confidence around an AI-based decision system of breast nodule assessment in 2D ultrasound
Meng et al. A cervical histopathology dataset for computer aided diagnosis of precancerous lesions
CN111798425A (zh) 基于深度学习的胃肠道间质瘤中核分裂象智能检测方法
Li et al. An MSCT-based radiomics nomogram combined with clinical factors can identify Crohn’s disease and ulcerative colitis
Ruan et al. Development and validation of a deep neural network for accurate identification of endoscopic images from patients with ulcerative colitis and Crohn's disease
Kaliyugarasan et al. Pulmonary nodule classification in lung cancer from 3D thoracic CT scans using fastai and MONAI
Leichter et al. Improved mammographic interpretation of masses using computer-aided diagnosis
WO2016085236A1 (fr) Méthode et système de détermination automatique d'un cancer de la thyroïde
Jing et al. A comprehensive survey of intestine histopathological image analysis using machine vision approaches
WO2022139321A1 (fr) Système de rapport d'image de pathologie du cancer de la prostate à base d'intelligence artificielle
Alimu et al. Artificial intelligence with a deep learning network for the quantification and distinction of functional adrenal tumors based on contrast-enhanced CT images
Kapruwan et al. Artificial Intelligence Enabled Diagnostic Digital Cytopathology System for Cervical Intraepithelial Neoplasia Detection: Advantages and Challenges
Parkavi et al. Deep learning techniques for the detection and classification of oral cancer using histopathological images
Craine et al. A clinical system for digital imaging colposcopy
Huang et al. External validation based on transfer learning for diagnosing atelectasis using portable chest X-rays
KR102684355B1 (ko) 인공지능 기반의 전립선암 병리 영상 레포트 시스템
Maftouni et al. A mask-guided attention deep learning model for COVID-19 diagnosis based on an integrated CT scan images database
Giger et al. Results of an observer study with an intelligent mammographic workstation for CAD
Kesavamurthy et al. Pattern Classification using imaging techniques for Infarct and Hemorrhage Identification in the Human Brain
Lee et al. Learning radiologist’s step-by-step skill for cervical spinal injury examination: Line drawing, prevertebral soft tissue thickness measurement, and swelling detection
Sun et al. CAMELYON 17 Challenge: A Comparison of Traditional Machine Learning (SVM) with the Deep Learning Method

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: 21911388

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: 21911388

Country of ref document: EP

Kind code of ref document: A1