Classifications

• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T7/00—Image analysis
  • G06T7/0002—Inspection of images, e.g. flaw detection
  • G06T7/0012—Biomedical image inspection
• • 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
  • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
  • G16H20/10—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
• • 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
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/10—Image acquisition modality
  • G06T2207/10056—Microscopic image
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/20—Special algorithmic details
  • G06T2207/20081—Training; Learning
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/20—Special algorithmic details
  • G06T2207/20084—Artificial neural networks [ANN]
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/30—Subject of image; Context of image processing
  • G06T2207/30004—Biomedical image processing
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/30—Subject of image; Context of image processing
  • G06T2207/30004—Biomedical image processing
  • G06T2207/30024—Cell structures in vitro; Tissue sections in vitro
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/30—Subject of image; Context of image processing
  • G06T2207/30004—Biomedical image processing
  • G06T2207/30056—Liver; Hepatic

Definitions

• the technology of the present disclosure relates to an information processing device, an operating method for the information processing device, and an operating program for the information processing device.
• Patent No. 6407242 describes the magnitude of change caused in a living body by a drug candidate substance (referred to as a "therapeutic compound” in Patent No. 6407242), etc., using a causal relationship that shows the causal relationship between multiple events representing changes in the living body.
• a system is described that performs quantitative calculations based on relational information (described as "network model” in Patent No. 6407242).
• a candidate substance is administered to a subject such as a rat.
• a pathologist identifies changes that occur in the subject after administration of the candidate substance.
• the pathologist estimates the mechanism of action of the candidate substance to produce drug efficacy or toxicity.
• the change that occurs in the subject after administration of the candidate substance is, for example, inflammation or morphological abnormality such as cancer that occurs in the organ of the subject.
• a pathologist identifies morphological abnormalities by observing a specimen image showing a tissue specimen of a subject's organ.
• One embodiment of the technology of the present disclosure provides an information processing device, an operating method of the information processing device, and an information processing device that can reduce the burden on pathologists and efficiently evaluate drug candidate substances. Provides an operating program.
• the information processing device of the present disclosure includes a processor, and the processor stores a plurality of pieces of causal relationship information indicating a causal relationship between a plurality of events representing changes in a living body, and a plurality of pieces of causal relationship information that are generated in a living subject to which a drug candidate substance has been administered.
• the system compares the generated event with an event that represents a change, extracts corresponding causal relationship information that includes an event that corresponds to the generated event from among a plurality of pieces of causal relationship information, and performs control to output the corresponding causal relationship information.
• the causal relationship information is information indicating a causal relationship between multiple events representing changes in a living body due to the toxicity of a past drug candidate substance.
• the occurring event represents a morphological abnormality that occurs in the organ of the subject after administration of the candidate substance.
• the processor identifies the type of morphological abnormality by performing an image analysis of a specimen image showing a tissue specimen of the subject's organ.
• the processor identifies multiple types of morphological abnormalities for one specimen image.
• the processor preferably uses a machine learning model that outputs the presence or absence of morphological abnormality in response to input of the specimen image.
• the processor extracts a plurality of corresponding causal relationship information
• the reliability is higher for corresponding causal relationship information that has a larger number of events that correspond to occurrence events.
• the occurrence event represents a morphological abnormality that occurred in the organ of the subject after administration of the candidate substance
• the processor outputs the presence or absence of the occurrence of the morphological abnormality in response to the input of a specimen image showing a tissue specimen of the organ of the subject. It is preferable to identify the type of morphological abnormality using a machine learning model, and derive the reliability based on the predicted probability of occurrence of the morphological abnormality by the machine learning model.
• the processor extracts a plurality of corresponding causal relationship information and a common event exists in at least two of the extracted plurality of corresponding causal relationship information, the processor extracts the corresponding causal relationship information in which the common event exists. It is preferable to control the output after integrating the events.
• Causal relationship information is prepared for each of multiple organs, and the processor outputs the corresponding causal relationship information in which a common event exists, after integrating the corresponding causal relationship information of different organs regarding the common event. Preferably, control is performed.
• the occurring event represents at least one of a change in the subject's body weight, a change in the subject's food intake, and a change in the result of a clinical chemistry test of the subject.
• the causal relationship information includes events related to changes in biological molecules, and when the processor performs control to output the corresponding causal relationship information, it is possible to distinguish events related to changes in biological molecules from other events. preferable.
• the operating method of the information processing device of the present disclosure includes a plurality of pieces of causal relationship information indicating a causal relationship between a plurality of events representing changes in a living body and a change that has occurred in a living subject to which a drug candidate substance has been administered. This includes comparing the information with the occurring event, extracting the corresponding causal relationship information including the event corresponding to the occurring event from among the multiple pieces of causal relationship information, and controlling the output of the corresponding causal relationship information. .
• the operating program of the information processing device of the present disclosure includes a plurality of pieces of causal relationship information indicating a causal relationship between a plurality of events representing changes in a living body, and a change that has occurred in a living subject to which a drug candidate substance has been administered. This includes comparing the information with the occurring event, extracting the corresponding causal relationship information including the event corresponding to the occurring event from among the multiple pieces of causal relationship information, and controlling the output of the corresponding causal relationship information. Have a computer perform a process.
• an information processing device an operating method for the information processing device, and an operating program for the information processing device are provided that can reduce the burden on pathologists and efficiently evaluate drug candidate substances. can be provided.
• FIG. 2 is a block diagram showing a computer that constitutes an information processing device.
• FIG. 2 is a block diagram showing a processing unit of a CPU of the information processing device.
• FIG. 3 is a diagram showing a morphological abnormality identification model. It is a figure which shows the patch image which subdivided the target specimen image in which the liver specimen was photographed.
• FIG. 7 is a diagram showing a patch image obtained by subdividing a target specimen image in which a pancreas specimen is photographed. It is a figure showing a hyperplasia identification model.
• FIG. 1 is a block diagram showing a computer that constitutes an information processing device.
• FIG. 2 is a block diagram showing a processing unit of a CPU of the information processing device.
• FIG. 3 is a diagram showing a morphological abnormality identification model. It is a figure which shows the patch image which subdivided the target specimen image in which the liver specimen was photographed.
• FIG. 7 is a diagram showing a patch image obtained by subdividing
• FIG. 6 is a diagram showing how an aggregated identification result is generated by outputting a identification result from each specific model in response to input of a patch image and aggregating each identification result.
• FIG. 6 is a diagram illustrating how occurrence event information is generated by further aggregating the aggregation identification results of each patch image.
• FIG. 3 is a diagram showing causal relationship information.
• FIG. 3 is a diagram showing first liver causality information.
• FIG. 3 is a diagram showing first pancreatic causal relationship information. It is a figure which shows the process of a collation extraction part. It is a figure which shows the process of a collation extraction part. It is a figure showing a target designation screen.
• FIG. 3 is a diagram showing an analysis result display screen.
• FIG. 3 is a flowchart showing a processing procedure of the information processing device. It is a table showing the number and reliability of events corresponding to morphological abnormalities.
• FIG. 7 is a diagram showing a mode in which a plurality of pieces of liver causality information extracted as corresponding causality information are displayed in descending order of reliability.
• FIG. 7 is a diagram showing a mode in which liver causality information whose reliability is equal to or higher than a set value is displayed in descending order of reliability among a plurality of pieces of liver causality information extracted as corresponding causality information. It is a table showing the average value and reliability of predicted probabilities of morphological abnormalities corresponding to events.
• FIG. 7 is a diagram showing first liver causal relationship information and first pancreatic causal relationship information in which a common event exists.
• FIG. 23 is a diagram showing integrated causal relationship information in which the first liver causal relationship information and the first pancreatic causal relationship information shown in FIG. 22 are integrated in a common event.
• FIG. 7 is a diagram showing an analysis result display screen displaying integrated causal relationship information.
• FIG. 3 is a diagram showing occurrence event information including a change in the subject's body weight, a change in the subject's food intake, and a change in the result of a clinical chemistry test of the subject.
• FIG. 3 is a diagram illustrating a mode of handling specimen images obtained by photographing a slide specimen on which tissue specimens of multiple types of organs are mounted.
• an information processing device 10 of the present disclosure is used to evaluate the efficacy and toxicity of a drug candidate substance 11.
• the information processing device 10 is, for example, a desktop personal computer, and includes a display 12 that displays various screens, and an input device 13 such as a keyboard, a mouse, a touch panel, and/or a microphone for voice input.
• the information processing device 10 is installed, for example, in a drug development facility, and is operated by a pathologist PT who is involved in drug development at the drug development facility.
• a specimen image 15 is input to the information processing device 10.
• the specimen image 15 is an image for evaluating the medicinal efficacy and toxicity of the candidate substance 11.
• the specimen image 15 is generated, for example, by the following procedure.
• a living subject S such as a rat
• a tissue sample of an organ of the subject S here a tissue sample of a cross section of the liver LV (hereinafter referred to as a liver sample).
• pancreatic specimen) LVS and a tissue specimen of a cross section of pancreatic PC hereinafter referred to as pancreatic specimen
• the collected liver specimen LVS and pancreas specimen PCS are pasted on a slide glass 16, and then the liver specimen LVS and pancreas specimen PCS are stained, here, with hematoxylin and eosin dye. Subsequently, the stained liver specimen LVS and pancreas specimen PCS are covered with a cover glass 17 to complete a slide specimen 18. Then, the slide specimen 18 is set on a photographing device 19 such as a digital optical microscope, and the specimen image 15 is photographed by the photographing device 19.
• the specimen image 15 obtained in this way is given a subject ID (Identification Data) for uniquely identifying the subject S, a specimen image ID for uniquely identifying the specimen image 15, and the date and time of photographing. .
• specimen image 15 in which the liver specimen LVS is photographed is referred to as a specimen image 15L
• specimen image 15 in which the pancreas specimen PCS is photographed is referred to as a specimen image 15P.
• specimen image 15L and 15P are collectively referred to as specimen image 15.
• one liver specimen LVS and one pancreas specimen PCS are taken, and only one specimen image 15L and 15P are taken each, but in reality, multiple images are taken from one subject S.
• a cross-sectional liver specimen LVS and a pancreas specimen PCS are collected, and a plurality of specimen images 15L and 15P are also photographed.
• a tissue specimen is also called a tissue section.
• the staining may be performed by staining with hematoxylin dye alone, staining with nuclear fast red dye, or the like.
• the administration group is composed of a plurality of subjects S to whom the candidate substance 11 has been administered.
• the control group is comprised of a plurality of subjects S to whom the candidate substance 11 was not administered, contrary to the administration group.
• an image of the liver specimen LVS and pancreas specimen PCS of the subject S in the administration group is used as the specimen image 15.
• the number of subjects S constituting the administration group and the number of subjects S constituting the control group are both, for example, about 5 to 10.
• the subjects S constituting the administration group and the subjects S constituting the control group have the same attributes and are placed under the same breeding environment.
• the same attributes include, for example, the same age in weeks and/or the same gender.
• the same attribute also includes the same age composition ratio and/or the same sex composition ratio (for example, 5 males and 5 females).
• the same breeding environment means, for example, that the feed is the same, the temperature and humidity of the breeding space are the same, and/or the size of the breeding space is the same.
• “Same” in the same rearing environment means not only the exact same, but also includes errors that are generally allowed in the technical field to which the technology of the present disclosure belongs and that do not go against the spirit of the technology of the present disclosure. Refers to the same meaning.
• the dosage of the candidate substance 11 is different.
• the dosage of the candidate substance 11 is varied in three stages: high dosage group, medium dosage group, and small dosage group. In this way, the influence on the subject S due to the dose of the candidate substance 11 can be determined.
• the computer constituting the information processing apparatus 10 includes, in addition to the display 12 and input device 13 described above, a storage 30, a memory 31, a CPU (Central Processing Unit) 32, and a communication unit 33. We are prepared. These are interconnected via a bus line 34.
• a bus line 34 The bus line 34.
• the storage 30 is a hard disk drive built into a computer that constitutes the information processing device 10 or connected through a cable or a network.
• the storage 30 is a disk array in which a plurality of hard disk drives are connected in series.
• the storage 30 stores control programs such as an operating system, various application programs, and various data accompanying these programs. Note that a solid state drive may be used instead of the hard disk drive.
• the memory 31 is a work memory for the CPU 32 to execute processing.
• the CPU 32 loads the program stored in the storage 30 into the memory 31 and executes processing according to the program. Thereby, the CPU 32 centrally controls each part of the computer.
• the CPU 32 is an example of a "processor" according to the technology of the present disclosure. Note that the memory 31 may be built into the CPU 32.
• the communication unit 33 controls transmission of various information to and from external devices such as the photographing device 19.
• an operating program 40 is stored in the storage 30 of the information processing device 10.
• the operating program 40 is an application program for causing the computer to function as the information processing device 10. That is, the operating program 40 is an example of "an operating program for an information processing device" according to the technology of the present disclosure.
• the storage 30 also stores a morphological abnormality identification model 41, causal relationship information 42, and the like.
• the morphological abnormality identification model 41 is an example of a "machine learning model” according to the technology of the present disclosure.
• the causal relationship information 42 indicates the causal relationship between a plurality of events 85 (see FIG. 11, etc.) representing changes in the living body.
• the term “living body” refers to a broadly general concept of "living thing” that is not limited to the subject S. Therefore, the term “living body” includes animals of a different species from the subject S.
• the CPU 32 of the computer constituting the information processing device 10 cooperates with the memory 31 and the like to control the read/write (hereinafter abbreviated as RW) control unit 50 and the abnormality identification.
• the unit 51 functions as a collation/extraction unit 52 , a display control unit 53 , and an instruction reception unit 54 .
• the RW control unit 50 controls storage of various data in the storage 30 and reading of various data in the storage 30.
• the RW control unit 50 stores the specimen image 15 from the imaging device 19 in the storage 30. Note that since a plurality of specimen images 15 are actually obtained from one subject S, a plurality of specimen images 15 for one subject S are stored in the storage 30.
• the RW control unit 50 reads out the specimen image 15 from the storage 30 according to the specification by the pathologist PT through the input device 13.
• the RW control section 50 outputs the read specimen image 15 to the morphological abnormality identification section 51 and the display control section 53.
• the specimen image 15 outputted from the RW control unit 50 to the morphological abnormality identifying unit 51 and the like is used to identify the type of morphological abnormality occurring in the liver specimen LVS and the pancreas specimen PCS.
• the target specimen image 15 for specifying the type of morphological abnormality occurring in the liver specimen LVS and the pancreas specimen PCS will be referred to as a target specimen image 15T.
• the RW control unit 50 reads the morphological abnormality identification model 41 from the storage 30 and outputs the read morphological abnormality identification model 41 to the morphological abnormality identification unit 51. Further, the RW control unit 50 reads the causal relationship information 42 from the storage 30 and outputs the read causal relationship information 42 to the matching extraction unit 52.
• the morphological abnormality identification unit 51 uses the morphological abnormality identification model 41 to identify the type of morphological abnormality that has occurred in the tissue specimens (liver specimen LVS and pancreas specimen PCS) shown in the target specimen image 15T.
• Morphological abnormalities include lesions that are not seen in normal tissue specimens, such as hyperplasia, infiltration, stasis, inflammation, tumor, canceration, proliferation, hemorrhage, and glycogen reduction.
• the morphological abnormality identification unit 51 outputs occurrence event information 60 that includes the identified type of morphological abnormality as an occurrence event to the collation extraction unit 52.
• the occurrence event represents a change that occurs in the subject S to whom the candidate substance 11 has been administered. In this example, the occurring event represents a morphological abnormality.
• the collation extraction unit 52 collates the event 85 of the causal relationship information 42 from the RW control unit 50 and the abnormal form of the occurrence event information 60. Then, the causal relationship information 42 including the event 85 that corresponds to the form abnormality in the occurrence event information 60 is extracted as the corresponding causal relationship information 61.
• the matching extraction unit 52 outputs the corresponding causal relationship information 61 to the display control unit 53.
• the display control unit 53 controls displaying various screens on the display 12.
• the various screens include a target designation screen 90 (see FIG. 15) for designating the target specimen image 15T, an analysis result display screen 95 (see FIG. 16) for displaying the target specimen image 15T and corresponding causal relationship information 61, etc. .
• the instruction receiving unit 54 receives various instructions from the pathologist PT via the input device 13 on various screens.
• the morphological abnormality identification model 41 includes a liver morphological abnormality identification model 41L and a pancreatic morphological abnormality identification model 41P.
• the liver morphological abnormality identification model 41L and the pancreatic morphological abnormality identification model 41P further include a plurality of specific models for each morphological abnormality. That is, the liver morphological abnormality identification model 41L includes a hyperplasia identification model 41L1 for identifying hyperplasia, a stasis identification model 41L2 for identifying stasis, an inflammation identification model 41L3 for identifying inflammation, and an inflammation identification model 41L3 for identifying canceration.
• the pancreatic morphological abnormality identification model 41P includes a hyperplasia identification model 41P1 for identifying hyperplasia, a stasis identification model 41P2 for identifying stasis, an inflammation identification model 41P3 for identifying inflammation, and an inflammation identification model 41P3 for identifying canceration. It has a cancerization specific model 41P4 for specifying cancer, and a proliferation specific model 41P5 for specifying proliferation.
• the morphological abnormality identification model 41 includes an infiltration identification model for identifying infiltration, a bleeding identification model for identifying bleeding, and the like.
• the morphological abnormality identifying unit 51 uses a well-known image recognition technique to identify the liver specimen LVS of a target specimen image 15T (hereinafter referred to as target specimen image 15LT) in which the liver specimen LVS is captured. is recognized, and the recognized liver specimen LVS is subdivided into a plurality of patch images 65LT.
• the morphological abnormality identifying unit 51 uses a well-known image recognition technique to generate a target specimen image 15T (hereinafter referred to as target specimen image 15PT) in which the pancreas specimen PCS is photographed.
• the pancreas specimen PCS is recognized, and the recognized pancreas specimen PCS is subdivided into a plurality of patch images 65PT.
• Patch images 65LT and 65PT have preset sizes that can be handled by liver morphological abnormality identification model 41L and pancreatic morphological abnormality identification model 41P.
• the morphological abnormality identification unit 51 assigns patch image IDs to patch images 65LT and 65PT.
• the morphological abnormality identifying unit 51 uses information indicating which position of the target specimen images 15LT and 15PT the patch images 65LT and 65PT are extracted from, that is, the position information of the patch images 65LT and 65PT, as the patch image ID. Associate. Note that in FIGS. 5 and 6, patch images 65LT and 65PT do not have areas that overlap with other patch images 65LT and 65PT, but patch images 65LT and 65PT partially overlap with other patch images 65LT and 65PT. You may do so.
• the hyperplasia identification model 41L1 includes an encoder section 70, a decoder section 71, a calculation section 72, and an output section 73.
• a patch image 65LT is input to the encoder section 70.
• the encoder unit 70 converts the patch image 65LT into a feature amount 74.
• the encoder section 70 passes the feature amount 74 to the decoder section 71.
• the decoder unit 71 decodes the feature amount 74.
• the encoder unit 70 includes a convolution layer that performs convolution processing using a filter, a pooling layer that performs pooling processing such as maximum value pooling processing, and the like.
• the hyperplasia identification model 41L1 is a convolutional neural network (CNN).
• the encoder unit 70 extracts the feature amount 74 by repeating the convolution process using the convolution layer and the pooling process using the pooling layer multiple times on the input patch image 65LT.
• the feature amount 74 represents the shape and texture characteristics of the liver specimen LVS shown in the patch image 65LT.
• the feature quantity 74 is a set of a plurality of numerical values. In other words, the feature amount 74 is multidimensional data. The number of dimensions of the feature amount 74 is, for example, 512, 1024, or 2048.
• the calculation unit 72 calculates a predicted probability 75 of the occurrence or non-occurrence of hyperplasia in the liver sample LVS shown in the patch image 65LT.
• the predicted probability 75 is, for example, a value between 0 and 1.0 (0% and 100%), and the closer the value is to 1.0, the more hyperplasia has occurred in the liver specimen LVS shown in the patch image 65LT. The probability is high.
• the calculation unit 72 outputs the predicted probability 75 to the output unit 73.
• the output unit 73 outputs a specific result 76L1 according to the predicted probability 75. More specifically, the output unit 73 compares the predicted probability 75 with a preset threshold. If the predicted probability 75 is less than the threshold, the output unit 73 outputs the identification result 76L1 indicating that hyperplasia has not occurred in the liver specimen LVS shown in the patch image 65LT (indicated as "no hyperplasia” in FIG. 7). Output. On the other hand, if the predicted probability 75 is equal to or greater than the threshold, the output unit 73 outputs the identification result indicating that hyperplasia has occurred in the liver specimen LVS shown in the patch image 65LT (indicated as "hyperplasia present" in FIG. 7). 76L1 is output.
• the threshold value is, for example, 0.5.
• other morphological abnormality identification models 41 such as the stasis identification model 41L2, the canceration identification model 41L4, the hyperplasia identification model 41P1, and the inflammation identification model 41P3 differ only in the content of the output identification result 76L. It has the same configuration as the specific model 41L1. Therefore, the hyperplasia identification model 41L1 will be described as a representative, and the description of the other morphological abnormality identification models 41 will be omitted.
• the morphological abnormality identification unit 51 stores one patch image 65LT in all the models constituting the liver morphological abnormality identification model 41L (hyperplasia identification model 41L1, stasis identification model 41L2, inflammation identification model). 41L3, cancerization specific model 41L4, proliferation specific model 41L5, etc.). Then, each model outputs a specific result 76L. Specifically, the hyperplasia identification model 41L1 outputs the identification result 76L1, the stasis identification model 41L2 outputs the identification result 76L2, and the inflammation identification model 41L3 outputs the identification result 76L3. Further, the cancerization identification model 41L4 outputs the identification result 76L4, and the proliferation identification model 41L5 outputs the identification result 76L5.
• the morphological abnormality identification unit 51 generates an aggregated identification result 80L by aggregating the plurality of identification results 76L outputted from each model in this way.
• the aggregated identification result 80L is a selection of the identification result 76 indicating that a morphological abnormality has occurred in the liver specimen LVS shown in the patch image 65LT, from among the plurality of identification results 76L output from each model.
• the specification result 76L2 outputted from the stasis specification model 41L2 has the content "stasis present”
• the specification result 76L3 output from the inflammation specification model 41L3 has the content "inflammation present”
• the other models A case is illustrated in which all of the identification results 76L outputted from the screen indicate "no abnormality in shape”.
• the aggregated identification result 80L has contents such as "inflammation” and "stasis” as shown in the figure. In this way, multiple types of morphological abnormalities may be identified for one patch image 65LT.
• the morphological abnormality identification unit 51 generates occurrence event information 60L by further aggregating the aggregation identification results 80L of each patch image 65LT.
• the occurrence event information 60L is a summary of the abnormalities that are determined to have occurred in the aggregated identification results 80L.
• FIG. 9 illustrates occurrence event information 60L that includes two morphological abnormalities, "inflammation” and "stasis", as occurrence events. In this way, multiple types of morphological abnormalities may be identified for one target specimen image 15LT.
• the morphological abnormality identification unit 51 generates occurrence event information 60P (see FIG. 14) by performing the same processing on the patch image 65PT as on the patch image 65LT.
• the causal relationship information 42 includes liver causal relationship information 42L and pancreatic causal relationship information 42P.
• the liver causality information 42L includes first liver causality information 42L1, second liver causality information 42L2, third liver causality information 42L3, fourth liver causality information 42L4, and fifth liver causality information 42L5.
• the pancreatic causal relationship information 42P includes first pancreatic causal relationship information 42P1, second pancreatic causal relationship information 42P2, third pancreatic causal relationship information 42P3, fourth pancreatic causal relationship information 42P4, fifth pancreatic causal relationship information 42P5, etc. has. In this way, the causal relationship information 42 is prepared for each of a plurality of organs.
• the first liver causal relationship information 42L1 and the first pancreatic causal relationship information 42P1 include a plurality of events 85 representing changes in the living body. Events 85 are connected by arrows. An event 85 connected to the start point of the arrow indicates the cause, and an event 85 connected to the end point of the arrow indicates the effect. Each event 85 is classified as a change at the molecular level, a change at the cellular level, or a change at the tissue level.
• the first liver causality information 42L1 starts with an event 85 of "decreased bile acid excretion” which is a change at the molecular level of the living body, and an event 85 of "inflammation-inducing substance release” which is also a change at the molecular level.
• This is information showing the causal relationship between each event 85, which is connected to the event 85 of "inflammation enhancement” which is a change at the cellular level, and finally leads to the event 85 of "cholestasis” which is a change at the tissue level. That is, the first liver causality information 42L1 is information indicating the causality of a plurality of events 85 representing changes in the living body due to the toxicity of past drug candidate substances.
• second liver causality information 42L2 Although not shown, the same applies to other second liver causality information 42L2, third liver causality information 42L3, etc.
• event 85 of “decreased bile acid excretion” and the event 85 of “release of inflammation-inducing substance” are examples of “events related to changes in biological molecules” according to the technology of the present disclosure.
• the first pancreatic causal relationship information 42P1 starts with an event 85 of "decreased lipid degradation” which is a change at the biological molecular level, an event 85 of "cholestasis” which is a change at the tissue level, and an event 85 of "cholestasis” which is a change at the tissue level.
• an event 85 of "decreased lipid degradation” which is a change at the biological molecular level
• an event 85 of "cholestasis” which is a change at the tissue level
• an event 85 of "cholestasis” which is a change at the tissue level.
• the first pancreatic causality information 42P1 is also information indicating the causality of a plurality of events 85 representing changes in the living body due to the toxicity of past drug candidate substances.
• the other second pancreatic causality information 42P2, third pancreatic causality information 42P3, etc. are similar.
• the event 85 of "decreased lipolysis” is an example of the "event related to changes in biological molecules" according to the technology of the present disclosure.
• the first pancreatic causality information 42P1 also includes some events 85 related to the liver LV. However, since the final event 85 is "pancreatic cancer" and is related to pancreatic PC, it is classified as pancreatic causality information 42P.
• Such causal relationship information 42 may be created by the pathologist PT based on the results of past evaluation tests of drug candidate substances. Furthermore, the causal relationship information 42 may be obtained from a public database such as AOP (Advance Outcome Pathway)-Wiki. Note that the causal relationship information 42 does not have to indicate the relationship between the events 85 with arrows as in the example, but may simply have information on the events 85 only.
• AOP Advanced Outcome Pathway
• the matching extraction unit 52 matches the causal relationship information 42 and the occurrence event information 60.
• FIG. 13 shows how the first liver causality information 42L1 and the occurrence event information 60L are compared.
• the verification extraction unit 52 identifies that the "inflammation enhancement" event 85 of the first liver causality information 42L1 corresponds to the "inflammation” of the occurrence event information 60L.
• the verification extraction unit 52 identifies, through verification, that the event 85 of "cholestasis" in the first liver causal relationship information 42L1 corresponds to "stasis" in the occurrence event information 60L.
• the collation extraction unit 52 uses the first liver causal relationship information 42L1 as the corresponding causal relationship information 61. Extract. Note that, although not shown, the verification extraction unit 52 naturally does not extract the liver causality information 42L that does not include the event 85 that corresponds to the morphological abnormality in the occurrence event information 60L as the corresponding causality information 61.
• FIG. 14 shows how the first pancreatic causal relationship information 42P1 and the occurrence event information 60P are compared.
• the verification extraction unit 52 identifies, through verification, that the event 85 of "pancreatic cell proliferation" in the first pancreatic causal relationship information 42P1 corresponds to "proliferation” in the occurrence event information 60P. Further, the verification extraction unit 52 identifies, through verification, that the event 85 of "pancreatic cancer” in the first pancreatic causal relationship information 42P1 corresponds to "cancerization” in the occurrence event information 60P.
• the collation extraction unit 52 uses the first pancreatic causal relationship information 42P1 as the corresponding causal relationship information 61. Extract. Although not shown, the collation and extraction unit 52 naturally does not extract the pancreatic causal relationship information 42P that does not include the event 85 that corresponds to the morphological abnormality in the occurrence event information 60P as the corresponding causal relationship information 61.
• the display control unit 53 controls displaying a target designation screen 90 shown in FIG. 15 on the display 12, as an example.
• the display control unit 53 displays the target designation screen 90 when the pathologist PT issues an instruction to display the specimen image 15 through the input device 13 and the instruction reception unit 54 accepts the display instruction.
• a plurality of specimen images 15 are displayed side by side. These plurality of specimen images 15 are specimen images 15 obtained from one subject S among the plurality of subjects S constituting the administration group.
• five specimen images 15L with specimen image IDs "SIL00001" to "SIL00005" obtained from a subject S whose specimen ID is "R001" and specimen image IDs "SIP00001" to "SIP00005" are shown. This shows an example in which five specimen images 15P are displayed side by side.
• the subject ID is a pull-down menu 91, and it is possible to switch the subject S whose specimen image 15 is displayed on the target specification screen 90.
• the target designation screen 90 is a screen for designating one target specimen image 15LT from among the plurality of specimen images 15L. Further, the target designation screen 90 is a screen for designating one target specimen image 15PT from among the plurality of specimen images 15P.
• the target designation screen 90 is provided with one selection frame 92L that can be moved between each specimen image 15L and one selection frame 92P that can be moved between each specimen image 15P.
• an analysis button 93 is provided at the bottom of the target specification screen 90. The pathologist PT selects the analysis button 93 after aligning the selection frames 92L and 92P with the desired specimen images 15L and 15P.
• the specimen images 15L and 15P in which the selection frames 92L and 92P are combined are set as the target specimen images 15LT and 15PT, the morphological abnormality identification unit 51 identifies the type of morphological abnormality, and the matching extraction unit 52 identifies the corresponding causal relationship information 61. Extraction etc. are performed.
• the display control unit 53 When the correspondence causality information 61 is input from the matching extraction unit 52, the display control unit 53 performs control to display an analysis result display screen 95 shown in FIG. 16 on the display 12, as an example.
• the analysis result display screen 95 displays target specimen images 15LT and 15PT, occurrence event information 60L and 60P (indicated as "morphological abnormality thought to have occurred" in FIG. 16), and corresponding causal relationship information 61. .
• the first liver causality information 42L1 shown in FIGS. 11 and 13 and the first pancreas causality information 42P1 shown in FIGS. 12 and 14 are displayed as the corresponding causality information 61. is exemplified.
• the control for displaying the analysis result display screen 95 including this correspondence causality information 61 on the display 12 is an example of "control for outputting correspondence causality information" according to the technology of the present disclosure.
• the display control unit 53 displays a highlighting frame 96 on an event 85 related to a change in a biological molecule among each event 85 in the corresponding causal relationship information 61 to distinguish it from other events 85 .
• the events 85 related to changes in biological molecules include both events 85 of "decreased bile acid excretion" and "release of inflammation-inducing substances” in the first liver causal relationship information 42L1, and the first pancreatic causal relationship information 42P1. This is event 85 of “decreased lipolysis”.
• the color of the event 85 related to a change in biological molecules is changed from the color of other events 85. Methods such as making the related event 85 larger in size than other events 85 or displaying the event 85 related to changes in biological molecules in a blinking manner may be adopted.
• the photographing device 19 photographs a specimen image 15L in which the liver specimen LVS of the subject S is photographed, and a specimen image 15P in which the pancreas specimen PCS of the subject S is photographed.
• Specimen images 15L and 15P are output from the imaging device 19 to the information processing device 10.
• the specimen images 15L and 15P from the imaging device 19 are stored in the storage 30 by the RW control unit 50.
• the RW control unit 50 reads out the specimen image 15 designated by the display instruction from the storage 30 (step ST100).
• the sample image 15 is output from the RW control section 50 to the display control section 53.
• the specimen image 15 is displayed on the display 12 through the target designation screen 90 under the control of the display control section 53 (step ST105).
• the RW control unit 50 causes the selection frame to be selected at that time.
• Specimen images 15L and 15P in which frames 92L and 92P have been aligned are read out from storage 30 as target specimen images 15LT and 15PT (step ST115).
• the target specimen images 15LT and 15PT are output from the RW control section 50 to the morphological abnormality identification section 51 and the display control section 53.
• the RW control unit 50 reads the morphological abnormality identification model 41 from the storage 30 and outputs the read morphological abnormality identification model 41 to the morphological abnormality identification unit 51. Further, the causal relationship information 42 is read out from the storage 30 by the RW control unit 50, and the read causal relationship information 42 is output to the matching extraction unit 52.
• the morphological abnormality identification unit 51 subdivides the target specimen images 15LT and 15PT into a plurality of patch images 65LT and 65PT (step ST120). Subsequently, as shown in FIG. 8, in the morphological abnormality identifying unit 51, the patch image 65LT is input to the hyperplasia identifying model 41L1, etc., and the identifying result 76L1, etc. is output from the hyperplasia identifying model 41L1, etc. Next, the morphological abnormality identification unit 51 aggregates the plurality of identification results 76L output from each model, and generates an aggregated identification result 80L. As a result, the type of morphological abnormality occurring in the liver specimen LVS shown in the patch image 65LT is specified (step ST125). The process of step ST125 is performed on all patch images 65LT. Similar processing is performed on patch image 65PT.
• step ST125 After the process in step ST125 is performed on all patch images 65LT and 65PT (YES in step ST130), as shown in FIG. Occurrence event information 60L is generated from this (step ST135). Similar processing is performed on patch image 65PT, and occurrence event information 60P is generated (step ST135). The occurrence event information 60L and 60P are output from the morphological abnormality identification section 51 to the collation extraction section 52.
• the causal relationship information 42 and the occurrence event information 60 are collated, and the corresponding causal relationship information 61 is extracted from the plurality of causal relationship information 42. (Step ST140).
• the correspondence causality information 61 is output from the matching extraction unit 52 to the display control unit 53.
• the analysis result display screen 95 shown in FIG. 16 is displayed on the display 12 (step ST145).
• the analysis result display screen 95 includes correspondence causality information 61.
• the pathologist PT views the analysis result display screen 95 and refers to the corresponding causal relationship information 61 and the like to estimate the mechanism of action of the toxicity of the candidate substance 11.
• the CPU 32 of the information processing device 10 includes the verification extraction section 52 and the display control section 53.
• the collation extraction unit 52 generates a plurality of pieces of causal relationship information 42 indicating the causal relationship between a plurality of events 85 representing changes in a living body, and occurrence information 42 representing a change that has occurred in a living subject S to which the drug candidate substance 11 has been administered. Compare with the event. Thereby, the matching extraction unit 52 extracts the corresponding causal relationship information 61 including the event 85 corresponding to the morphological abnormality from among the plurality of causal relationship information 42.
• the display control unit 53 performs control to display an analysis result display screen 95 including the correspondence causality information 61 on the display 12.
• the corresponding causal relationship information 61 is very useful for the pathologist PT to estimate the mechanism of action of the toxicity of the candidate substance 11. Therefore, it becomes possible to reduce the burden on the pathologist PT and efficiently evaluate the candidate substance 11.
• the causal relationship information 42 is information indicating the causal relationship between a plurality of events 85 representing changes in the living body due to the toxicity of past drug candidate substances. This facilitates the pathologist PT's estimation of the mechanism of action of the toxicity of the candidate substance 11. Note that instead of or in addition to the causal relationship information 42 indicating the causal relationship between a plurality of events 85 representing changes in the living body due to the toxicity of past drug candidate substances, Causal relationship information 42 indicating the causal relationship between the plurality of represented events 85 may be used.
• the occurrence event represents a morphological abnormality that occurred in the organs of the subject S, here the liver specimen LVS and the pancreas specimen PCS, after administration of the candidate substance 11.
• the event 85 of the causal relationship information 42 is a form, such as "hyperinflammation” and “cholestasis” of the first liver causal relationship information 42L1, or "pancreatic cell proliferation” and “pancreatic cancer” of the first pancreatic causal relationship information 42P1.
• the morphological abnormality identification unit 51 identifies the type of morphological abnormality by analyzing the target specimen images 15LT and 15PT in which the tissue specimens of the organs of the subject S, here the liver specimen LVS and the pancreas specimen PCS, are captured. Therefore, compared to the case where the pathologist PT specifies the type of morphological abnormality by observing the target specimen images 15LT and 15PT, the burden on the pathologist PT can be further reduced.
• the morphological abnormality identifying unit 51 identifies multiple types of morphological abnormalities for one target specimen image 15LT or 15PT. Therefore, compared to the case where only one type of morphological abnormality can always be identified for one target specimen image 15LT or 15PT, the corresponding causal relationship is more useful for estimating the mechanism of action of the toxicity expression of the candidate substance 11.
• Information 61 can be extracted.
• the morphological abnormality identification unit 51 uses a morphological abnormality identification model 41 that outputs the presence or absence of morphological abnormality in response to input of target specimen images 15LT and 15PT (patch images 65LT and 65PT).
• machine learning models such as the morphological abnormality identification model 41 with relatively high prediction accuracy can be easily prepared. Therefore, the type of morphological abnormality can be specified easily and accurately. Note that by using a well-known technique such as a pattern recognition technique, the type of morphological abnormality may be identified without using a machine learning model.
• the causality information 42 includes events 85 related to changes in biological molecules.
• the display control unit 53 displays a highlighting frame 96 on an event 85 related to a change in biological molecules, and displays a highlighting frame 96 on an event 85 related to a change in a biological molecule. Distinguish.
• the event 85 related to changes in biological molecules is considered to be a promising candidate for a pharmacodynamic biomarker that is important for estimating the mechanism of action of the toxicity of the candidate substance 11. Therefore, by distinguishing the event 85 related to a change in biological molecules from other events 85, the pathologist PT can be made aware of the event 85 related to a change in biological molecules.
• [2nd_1 embodiment] In the first embodiment described above, a case has been exemplified in which one piece of correspondence causality information 61 is extracted for each of the liver LV and pancreas PC, but the present invention is not limited to this.
• the matching extraction unit 52 may extract a plurality of pieces of corresponding causal relationship information 61.
• the display control unit 53 derives the reliability of each of the plurality of pieces of corresponding causal relationship information 61 based on the table 100 shown in FIG. 18, for example.
• the table 100 is data in which reliability levels are registered according to the number of events 85 (hereinafter referred to as the number of corresponding events) that correspond to the form abnormality in the occurrence event information 60. For example, when the number of corresponding events is 1, the reliability is 1, when the number of corresponding events is 4, the reliability is 4, and when the number of corresponding events is 5 or more, the reliability is 5. The higher the reliability value of the corresponding causal relationship information 61, the higher the reliability of the information. Taking as an example the first liver causality information 42L1 shown in FIG. 13 etc. and the first pancreas causality information 42P1 shown in FIG. 14 etc., the number of corresponding events is 2, so the reliability is 2. Become.
• the display control unit 53 displays the corresponding causal relationship information 61 in descending order of reliability on the analysis result display screen 95. do.
• five liver causality information 42L6, 7th liver causality information 42L7, 8th liver causality information 42L8, 9th liver causality information 42L9, and 10th liver causality information 42L10 are shown.
• a case is illustrated in which the relationship information 42L is extracted as the corresponding causal relationship information 61.
• the reliability of the sixth liver causality information 42L6 is 1
• the reliability of the seventh liver causality information 42L7 and the tenth liver causality information 42L10 is 2
• the reliability of the eighth liver causality information 42L8 is illustrated.
• the display control unit 53 sets the display order of the eighth liver causal relationship information 42L8, which has the highest reliability of 5, as the first on the analysis result display screen 95.
• the display order of the 9th liver causality information 42L9 will be number 2
• the display order of the 7th liver causality information 42L7 will be number 3
• the display order of the 10th liver causality information 42L10 will be number 4
• the reliability will be determined.
• the display order of the sixth liver causality information 42L6, which is the lowest in number 1, is set as number 5.
• the display control unit 53 displays a screen with a reliability level of 3 or more on the analysis result display screen 95.
• the corresponding causal relationship information 61 is displayed in descending order of reliability.
• the extracted correspondence causality information 61 and its reliability are the same as in the example of FIG. 19.
• the display control unit 53 sets the display order of the eighth liver causal relationship information 42L8, which has the highest reliability level of 5, first on the analysis result display screen 95, and sets the display order of the eighth liver causal relationship information 42L8, which has the highest reliability level of 5, as the first in the display order on the analysis result display screen 95, and sets the display order of the eighth liver causal relationship information 42L8, which has the highest reliability level of 5, as the 9th liver causal relationship information that has the second highest reliability level of 4.
• the display order of the information 42L9 is set as number 2. Sixth liver causality information 42L6, seventh liver causality information 42L7, and tenth liver causality information 42L10 whose reliability is 1 or 2 and less than the set value of 3 are hidden.
• the display control unit 53 derives the reliability of each of the plurality of pieces of corresponding causal relationship information 61. Then, control is performed to display the corresponding causal relationship information 61 based on the reliability. Therefore, as shown in FIG. 19, a plurality of pieces of corresponding causal relationship information 61 are displayed in descending order of reliability, or as shown in FIG. 20, corresponding causal relationship information 61 whose reliability is less than a set value is hidden. You can Estimation of the mechanism of action of the toxicity of candidate substance 11 by the pathologist PT will further progress.
• the reliability is higher as the corresponding causal relationship information 61 has a larger number of corresponding events. Therefore, the corresponding causal relationship information 61 with a large number of corresponding events can be displayed preferentially over other corresponding causal relationship information 61. Corresponding causal relationship information 61 with a large number of corresponding events is considered to be more useful for estimating the mechanism of action of toxicity of the candidate substance 11. Therefore, the pathologist PT can further estimate the mechanism of action of the toxicity of the candidate substance 11. Note that the analysis result display screen 95 may display reliability according to the number of corresponding events.
• the display control unit 53 determines the reliability of each of the plural pieces of corresponding causal relationship information 61 based on the table 105 shown in FIG. 21, for example. Derive the degree.
• Table 105 is data in which reliability levels are registered according to the average value of the predicted probabilities 75 of morphological abnormalities corresponding to the events 85 of the corresponding causal relationship information 61.
• the reliability is 1, and when the average value of predicted probabilities 75 is 0.7 or more and less than 0.8, the reliability is 3, and the predicted probabilities 75 are The reliability is 5 when the average value of is 0.9 or more and 1.0 or less.
• the average value of the predicted probability 75 is calculated as follows.
• the first liver causal relationship information 42L1 shown in FIG. 13 etc. will be taken as an example again.
• the predicted probability 75 by the inflammation identification model 41L3 of "inflammation” which is a morphological abnormality corresponding to the "inflammation enhancement” event 85 is 0.85
• "stasis” which is a morphological abnormality corresponding to the "cholestasis” event 85.
• the predicted probability 75 by the stagnation identification model 41L2 is, for example, 0.65.
• the predicted probability 75 of "inflammation” by the inflammation identification model 41L3 is a representative value of the predicted probability 75 obtained from all the patch images 65LT in which the identification result 76L3 indicating "inflammation” has been output.
• the predicted probability 75 of "stasis” by the stagnation identification model 41L2 is a representative value of the predicted probability 75 obtained from all the patch images 65LT for which the specification result 76L2 indicating "stagnation” has been output. Representative values include an average value, maximum value, median value, and minimum value.
• the display control unit 53 displays a plurality of pieces of corresponding causal relationship information 61 in descending order of reliability, and corresponding causal relationship information whose reliability is less than a set value, as in the case of the 2_1 embodiment. 61 may be hidden. Thereby, as in the case of the 2_1 embodiment, it is possible to obtain the effect that the pathologist PT can further estimate the mechanism of action of the toxicity of the candidate substance 11.
• the display control unit 53 derives the reliability based on the predicted probability 75 of the occurrence or non-occurrence of a morphological abnormality by the morphological abnormality identification model 41. Therefore, the corresponding causal relationship information 61 with a high average predicted probability 75 can be displayed preferentially over other corresponding causal relationship information 61. Correspondence causality information 61 with a high average predicted probability 75 is considered to be more useful for estimating the mechanism of action of toxicity of the candidate substance 11. Therefore, the pathologist PT can further estimate the mechanism of action of the toxicity of the candidate substance 11.
• the analysis result display screen 95 may display the reliability according to the average value of the predicted probabilities 75 or the average value of the predicted probabilities 75 itself. Further, instead of the average value of the predicted probability 75 of the morphological abnormality corresponding to the event 85 of the corresponding causal relationship information 61, the maximum value, median value, minimum value, etc. may be used.
• the 2_1 embodiment and the 2_2 embodiment may be implemented in combination.
• the display control unit 53 derives the reliability based on the data in which the reliability according to the average value of the number of corresponding events and the predicted probability 75 is registered.
• the display control unit 53 may derive the reliability by solving a calculation formula using the number of corresponding events and the average value of the predicted probabilities 75 as parameters.
• the reliability may be derived based on the ratio between the total number of events 85 that constitute the correspondence causal relationship information 61 and the number of corresponding events. In this case, for example, Fisher's exact test is performed, and the value obtained by subtracting the p value calculated thereby from 1 is used as the reliability level.
• the display control unit 53 integrates the corresponding causal relationship information 61 in which the common event 85 exists in the common event 85.
• FIG. 22 exemplifies a case where the collation extraction unit 52 extracts the first liver causal relationship information 42L1 and the first pancreatic causal relationship information 42P1 as the corresponding causal relationship information 61, as in the first embodiment.
• the first liver causal relationship information 42L1 and the first pancreatic causal relationship information 42P1 have the event 85 of "cholestasis” in common.
• the first liver causality information 42L1 and the first pancreas causality information 42P1 are examples of "corresponding causality information in which a common event exists" and "corresponding causality information of different organs” according to the technology of the present disclosure. It is.
• the event 85 of “cholestasis” is an example of a “common event” according to the technology of the present disclosure.
• the display control unit 53 integrates the first liver causal relationship information 42L1 and the first pancreatic causal relationship information 42P1 in the event 85 of "cholestasis", and creates an integrated causal relationship. Let it be information 110. Then, as shown in FIG. 24 as an example, the display control unit 53 performs control to display the integrated causal relationship information 110 on the analysis result display screen 95.
• the display control unit 53 allows the collation extraction unit 52 to extract a plurality of pieces of corresponding causal relationship information 61, and to display at least two of the plurality of extracted pieces of corresponding causal relationship information 61.
• control is performed to display the corresponding causal relationship information 61 in which the common event 85 exists, after integrating it in the common event 85. Therefore, the corresponding causal relationship information 61 in which the common event 85 exists can be summarized compactly, and the corresponding causal relationship information 61 in which the common event 85 exists can be displayed in an easy-to-read manner.
• the display control unit 53 integrates the corresponding causal relationship information 61 of different organs, the first liver causal relationship information 42L1 and the first pancreas causal relationship information 42P1. Therefore, the correspondence causality information 61 of different organs can be compactly summarized, and the correspondence causality information 61 of different organs can be displayed in an easy-to-read manner.
• the display control unit 53 derives reliability for the integrated causal relationship information 110.
• occurrence event information 115 shown in FIG. 25 shows changes in the body weight of the subject S after administration of the candidate substance 11, changes in the amount of food intake (feeding amount) of the subject S, and clinical chemistry of the subject S. It may also be an occurring event that represents a change in the result of a test. Changes in the body weight of the subject S are "weight loss” and “weight gain” shown in the figure. The changes in the amount of food eaten by the subject S are a "decreased amount of food eaten" and an "increase in the amount of food eaten” shown in the figure.
• changes in the clinical chemistry test results of the subject S are an "increase in blood sugar level” and a “decrease in blood sugar level” as shown in the figure.
• changes in the results of clinical chemistry tests of subject S include “increase in cholesterol level” and “decrease in cholesterol level”, “increase in uric acid level” and “decrease in uric acid level”, or “increase in ⁇ -GTP (Glutamyl Transpeptidase)”. ” and “ ⁇ -GTP decrease”.
• the occurrence event represents a change in the body weight of the subject S, a change in the amount of food intake of the subject S, and a change in the result of the clinical chemistry test of the subject S. Therefore, the causal relationship information 42 having events 85 related to changes in the body weight of the subject S, changes in the amount of food intake of the subject S, and changes in the results of clinical chemistry tests of the subject S is used as the corresponding causal relationship information 61. can be extracted. As shown in the occurrence event information 115 shown in FIG. 25, in addition to the occurrence event representing morphological abnormality, there are also changes in the body weight of the subject S, changes in the amount of food intake of the subject S, and results of clinical chemistry tests of the subject S. By adding occurrence events that represent changes, it is possible to increase the variations of the corresponding causal relationship information 61.
• occurrence events representing changes in the body weight of the subject S, changes in the amount of food intake of the subject S, and changes in the results of clinical chemistry tests of the subject S may be used.
• the change may be one or two of the changes in the body weight of the subject S, the change in the food intake of the subject S, and the changes in the clinical chemistry test results of the subject S, rather than all of them. .
• the fifth embodiment deals with a slide specimen 120 in which tissue specimens of a plurality of types of organs are placed on a single slide glass 16.
• FIG. 26 illustrates a case where a heart specimen HS, a brain specimen BS, and a bone marrow specimen BMS are placed in addition to the liver specimen LVS.
• the specimen image 15 includes a heart specimen HS, a brain specimen BS, and a bone marrow specimen BMS in addition to the liver specimen LVS.
• the CPU 32 of the information processing device 10 of the fifth embodiment functions as the identification unit 121 in addition to each of the processing units 50 to 54 of the first embodiment.
• the identification unit 121 identifies the tissue specimen of each organ from the specimen image 15 using, for example, a template for identifying the tissue specimen of each organ or a machine learning model.
• the identification unit 121 outputs coordinate information of frames 122 to 125 surrounding the tissue specimens of each organ as identification results.
• a frame 122 is a frame surrounding the heart sample HS
• a frame 123 is a frame surrounding the liver sample LVS.
• a frame 124 is a frame surrounding the brain sample BS
• a frame 125 is a frame surrounding the bone marrow sample BMS.
• the specimen image 15 is an image obtained by photographing the slide specimen 120 on which tissue specimens of multiple types of organs are mounted.
• the identification unit 121 identifies tissue specimens of each organ from such specimen images 15. Therefore, it is possible to handle slide specimens 120 on which tissue specimens of multiple types of organs are mounted.
• the slide specimen is not the slide specimen 18 on which a tissue specimen of one organ is placed as in the first embodiment, but rather the slide specimen 120 on which tissue specimens of multiple types of organs are placed as in the present embodiment. is more common. Therefore, it is possible to perform processing that is more suitable for general operations.
• a frame indicating the tissue specimen of each organ in the specimen image 15 may be manually defined by the pathologist PT.
• the portion where the morphological abnormality has occurred may be clearly indicated by coloring or the like.
• the output mode of the correspondence causality information 61 is not limited to the mode in which the exemplary analysis result display screen 95 is displayed on the display 12.
• the corresponding causal relationship information 61 may be printed out on a paper medium, or the corresponding causal relationship information 61 may be attached to an e-mail and sent.
• the organ is not limited to the illustrated liver LV, etc.
• the stomach, lungs, small intestine, large intestine, etc. may be used.
• the subject S is not limited to rats. It may also be a mouse, guinea pig, sand mouse, hamster, ferret, rabbit, dog, cat, or monkey.
• the information processing device 10 may be a personal computer installed in a pharmaceutical development facility as shown in FIG. 1, or a server computer installed in a data center independent from the pharmaceutical development facility.
• the specimen image 15 is sent from a personal computer installed at each drug development facility to the server computer via a network such as the Internet.
• the server computer delivers various screens such as the target specification screen 90 to the personal computer in the form of screen data for web distribution created using a markup language such as XML (Extensible Markup Language).
• the personal computer reproduces the screen displayed on the web browser based on the screen data and displays this on the display.
• JSON JavaScript (registered trademark) Object Notation
• JSON JavaScript (registered trademark) Object Notation
• the information processing device 10 can be widely used throughout all stages of drug development, from the initial stage of drug discovery target setting to the final stage of clinical trials.
• the hardware configuration of the computer configuring the information processing device 10 can be modified in various ways.
• the information processing device 10 may be configured with a plurality of computers separated as hardware for the purpose of improving processing power and reliability.
• the functions of the morphological abnormality identifying section 51 and the matching and extracting section 52 may be distributed between two computers.
• the information processing device 10 is composed of two computers.
• the hardware configuration of the computer of the information processing device 10 can be changed as appropriate depending on required performance such as processing capacity, safety, and reliability.
• application programs such as the operating program 40 can be duplicated or distributed and stored in multiple storages for the purpose of ensuring safety and reliability. .
• a processing unit that executes various processes, such as an RW control unit 50, a form abnormality identification unit 51, a collation extraction unit 52, a display control unit 53, an instruction reception unit 54, and an identification unit 121, is used.
• various processors include the CPU 32, which is a general-purpose processor that executes software (operating program 40) and functions as various processing units, as well as FPGA (Field Programmable Gate Array), etc.
• Dedicated processors are processors with circuit configurations specifically designed to execute specific processes, such as programmable logic devices (PLDs), which are processors whose circuit configurations can be changed, and ASICs (Application Specific Integrated Circuits). Includes electrical circuits, etc.
• PLDs programmable logic devices
• ASICs Application Specific Integrated Circuits
• One processing unit may be composed of one of these various processors, or a combination of two or more processors of the same type or different types (for example, a combination of multiple FPGAs and/or a CPU and (in combination with FPGA). Further, the plurality of processing units may be configured with one processor.
• one processor is configured with a combination of one or more CPUs and software, as typified by computers such as clients and servers.
• a processor functions as multiple processing units.
• SoC system-on-chip
• various processing units are configured using one or more of the various processors described above as a hardware structure.
• circuitry that is a combination of circuit elements such as semiconductor elements can be used.
• the processor includes: Compare multiple pieces of causal relationship information showing the causal relationships between multiple events representing changes in the living body with occurrence events representing changes that have occurred in the living subject to whom the drug candidate substance has been administered, extracting corresponding causal relationship information including the event corresponding to the occurring event from among the plurality of causal relationship information; performing control to output the corresponding causal relationship information; Information processing device.
• the information processing device according to appendix 1, wherein the causal relationship information is information indicating a causal relationship between a plurality of events representing changes in a living body due to toxicity of past drug candidate substances.
• the information processing device includes: The information processing device according to supplementary note 3, wherein the type of the morphological abnormality is identified by performing image analysis on a specimen image in which a tissue specimen of an organ of the subject is photographed.
• the processor includes: The information processing device according to supplementary note 4, wherein a plurality of types of the morphological abnormality are specified for one specimen image.
• the processor includes: The information processing apparatus according to additional item 4 or 5, which uses a machine learning model that outputs whether or not the morphological abnormality occurs in accordance with the input of the sample image in the image analysis.
• the processor includes: When multiple pieces of corresponding causal relationship information are extracted, Deriving the reliability of each of the plurality of corresponding causal relationship information, The information processing device according to any one of Supplementary Notes 1 to 6, which performs control to output the corresponding causal relationship information based on the reliability.
• the information processing device according to supplementary note 7, wherein the reliability is higher as the correspondence causal relationship information has a larger number of events that correspond to the occurring event.
• the occurrence event represents a morphological abnormality that occurs in the organ of the subject after administration of the candidate substance
• the processor includes: identifying the type of morphological abnormality using a machine learning model that outputs the type of morphological abnormality in response to input of a specimen image in which a tissue specimen of an organ of the subject is captured; The information processing device according to additional item 7 or 8, wherein the reliability is derived based on the predicted probability of occurrence of the morphological abnormality by the machine learning model.
• the processor includes: When a plurality of pieces of corresponding causal relationship information are extracted, and the event is common to at least two of the extracted plural pieces of corresponding causal relationship information, The information processing device according to any one of Supplementary Items 1 to 9, which performs control to output the corresponding causal relationship information in which the common event exists, after integrating the corresponding causal relationship information in the common event.
• the causal relationship information is prepared for each of multiple organs, The processor includes: The information processing according to supplementary note 10, which performs control to output the corresponding causal relationship information in which the common event exists, after integrating the corresponding causal relationship information of different organs in the common event. Device.
• the occurrence event is defined in Supplementary Notes 1 to 11, which represent at least one of a change in the body weight of the subject, a change in the amount of food intake of the subject, and a change in the results of a clinical chemistry test of the subject.
• the information processing device according to any one of the items.
• the causal relationship information includes an event related to a change in molecules of the living body,
• the processor includes: The information processing device according to any one of Supplementary Notes 1 to 12, which distinguishes events related to changes in molecules of the living body from other events when performing control to output the corresponding causal relationship information.
• a and/or B has the same meaning as “at least one of A and B.” That is, “A and/or B” means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, even when three or more items are expressed in conjunction with “and/or”, the same concept as “A and/or B" is applied.

Landscapes

• Engineering & Computer Science (AREA)
• Health & Medical Sciences (AREA)
• Medical Informatics (AREA)
• Public Health (AREA)
• General Health & Medical Sciences (AREA)
• Primary Health Care (AREA)
• Epidemiology (AREA)
• Data Mining & Analysis (AREA)
• Biomedical Technology (AREA)
• Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
• Radiology & Medical Imaging (AREA)
• Pathology (AREA)
• Databases & Information Systems (AREA)
• Bioinformatics & Cheminformatics (AREA)
• Chemical & Material Sciences (AREA)
• Medicinal Chemistry (AREA)
• Quality & Reliability (AREA)
• Computer Vision & Pattern Recognition (AREA)
• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Theoretical Computer Science (AREA)
• Investigating Or Analysing Biological Materials (AREA)
• Measuring And Recording Apparatus For Diagnosis (AREA)

Priority Applications (4)

Applications Claiming Priority (2)

Related Child Applications (1)

Publications (1)

Family

ID=89617796

Family Applications (1)

Country Status (5)

Citations (3)

Family Cites Families (1)

• 2023
  • 2023-07-19 WO PCT/JP2023/026382 patent/WO2024019081A1/ja not_active Ceased
  • 2023-07-19 CN CN202380055321.9A patent/CN119585801A/zh active Pending
  • 2023-07-19 JP JP2024535109A patent/JPWO2024019081A1/ja active Pending
  • 2023-07-19 EP EP23843004.5A patent/EP4560639A4/en active Pending
• 2025
  • 2025-01-21 US US19/033,053 patent/US20250166190A1/en active Pending

Patent Citations (3)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events