Drug research and development assistance device, method for operating drug research and development assistance device, and program for operating drug research and development assistance device

By extracting feature quantities of tissue section images of the Chinese medicine administration group and the control group of drug development and comparing machine learning models, the problem of abnormal judgment caused by drug development Chinese medicine candidate substances was solved, and the accuracy and reliability of abnormal judgment were improved. .

CN118946806APending Publication Date: 2024-11-12FUJIFILM CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380030385.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-03-29
Filing Date
2023-01-30
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In drug development, it is difficult to determine whether abnormalities caused by drug candidate substances are actually present in the subjects of the dosing group, especially because abnormalities caused by possible attributes and feeding environments cannot be accurately distinguished.

Method used

By obtaining tissue section images of the dosing group and the control group, feature quantities are extracted, and the distance and area ratios in the feature quantities space are calculated to determine whether there is an abnormality caused by the drug candidate substance.

Benefits of technology

It can effectively determine whether the candidate substance of the agent is true and leads to abnormalities in the subject, improves the accuracy of abnormal judgment in drug research and development, and reduces misjudgment caused by attributes and feeding environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118946806A_ABST
    Figure CN118946806A_ABST
Patent Text Reader

Abstract

This drug development assistance device is provided with a processor that performs: a process for acquiring a first slice image in which a tissue slice of at least one of a plurality of subjects constituting an administration group of a candidate substance to which a drug is administered is reflected; extracting a first feature quantity from the first slice image; a second slice image acquisition unit that acquires a plurality of second slice images in which tissue slices of a plurality of subjects constituting a control group of a candidate substance to which a drug is not administered, the plurality of subjects having the same properties as those of subjects constituting the administration group and being in the same feeding environment, and the plurality of subjects representing tissue slices of the plurality of subjects constituting the candidate substance to which the drug is not administered; extracting a second feature amount from each of the plurality of second slice images; acquiring third feature quantity information relating to third feature quantities extracted from each of a plurality of third slice images in which tissue slices of a plurality of subjects constituting a control group in which the candidate substance is not introduced in a past evaluation test are represented; and determining whether or not there is an abnormality in the tissue slice reflected in the first slice image on the basis of the first feature amount, the second feature amount, and the third feature amount information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The technology of the present invention relates to a drug development support device, a working method of the drug development support device, and a working procedure of the drug development support device. Background Art

[0002] In the field of drug development, tests are being conducted to evaluate the efficacy and toxicity of candidate substances by injecting them into rats and other test subjects. In such evaluation tests, slice images of tissue slices (brain slices, liver slices, heart slices, etc.) collected by autopsy of the test subject are used. In the past, pathologists observed the slice images for evaluation, but with the recent progress of image analysis technology, a technology has been developed that automatically performs evaluation without the need for pathologists to operate.

[0003] For example, the following is recorded in Japanese Patent Publication No. 2021-508373: Evaluation is performed using a machine learning model learned using slice images obtained by photographing tissue slices without abnormalities such as tumors and inflammation (recorded in Japanese Patent Publication No. 2021-508373 as "tissue samples with substantially no abnormalities"). Specifically, in Japanese Patent Publication No. 2021-508373, first, a slice image of a subject to which a candidate substance is injected is input into the above-mentioned machine learning model. Next, the difference between the feature quantity extracted by this extraction and the feature quantity extracted from the learning slice image obtained by photographing a tissue slice without abnormalities is calculated (recorded in Japanese Patent Publication No. 2021-508373 as "deviation from the normal score"). Then, based on the obtained difference, it is determined whether an abnormality occurs in the tissue slice of the subject to which the candidate substance is injected. Summary of the invention

[0004] Technical issues to be solved by the invention

[0005] The subjects are divided into a dosing group into which the candidate substance is administered and a control group into which the candidate substance is not administered for comparison with the dosing group. In the subjects of the control group, abnormalities will occur even if the candidate substance is not administered. It is believed that the abnormalities that occur in the subjects of the control group are due to attributes such as age and / or feeding environment such as feed. There are also many subjects in the dosing group that have abnormalities due to attributes and / or feeding environment. Therefore, when abnormalities occur in tissue sections of subjects in the dosing group, it is necessary to determine whether the cause of the abnormality is really the candidate substance. However, there is no record of abnormalities due to attributes and / or feeding environment in the Japanese Patent Gazette No. 2021-508373, and of course, there is no record of methods for determining whether the cause of the abnormality is a candidate substance.

[0006] One embodiment of the technology according to the present invention provides a drug development support device, a method for operating the drug development support device, and a program for operating the drug development support device that can determine whether an abnormality caused by a candidate substance of a drug occurs in a subject in a drug administration group.

[0007] Means for solving technical problems

[0008] The drug development support device of the present invention has a processor that performs the following processing: acquiring a first slice image, the first slice image showing a tissue slice of at least one subject among a plurality of subjects constituting a drug administration group of candidate substances to which a drug is administered; extracting a first feature quantity from the first slice image; acquiring a plurality of second slice images, the plurality of second slice images showing tissue slices of a plurality of subjects, the plurality of subjects having the same attributes as the subjects constituting the drug administration group and being in the same breeding environment, and constituting a control group of candidate substances not administered with the drug; extracting a second feature quantity from each of the plurality of second slice images; acquiring a third feature quantity information related to a third feature quantity, the third feature quantity being extracted from each of a plurality of third slice images showing tissue slices of a plurality of subjects constituting a control group to which the candidate substances were not administered in a past evaluation test; and determining whether there is an abnormality in the tissue slice shown in the first slice image based on the first feature quantity, the second feature quantity, and the third feature quantity information.

[0009] Preferably, the processor extracts the first feature amount and the second feature amount by using a machine learning model learned using a plurality of third slice images.

[0010] Preferably, the first feature quantity, the second feature quantity and the third feature quantity have the same dimension and can be compared in the same feature quantity space. The third feature quantity information is the information of the representative positions of multiple third feature quantities in the feature quantity space. The processor performs the following processing: calculates the first distance between the representative position in the feature quantity space and the position of the first feature quantity, calculates the second distance between the representative position in the feature quantity space and the respective positions of multiple second feature quantities, and makes a judgment based on the first distance and the second distance.

[0011] Preferably, the processor performs the following processing: extracting the first feature quantity for each of the plurality of first block images obtained by subdividing the first slice image, and calculating the first distance; extracting the second feature quantity for each of the plurality of second block images obtained by subdividing the second slice image, and calculating the second distance; and making a determination based on a first area ratio occupied by the first block images whose first distance is outside a set range in the first slice image and a second area ratio occupied by the second block images whose second distance is outside the set range in the second slice image.

[0012] Preferably, the processor determines that, when there is a statistically significant difference between the first area ratio and the second area ratio, an abnormality exists in the tissue slice shown in the first slice image.

[0013] Preferably, the processor determines that an abnormality exists in the tissue slice that appears in the first slice image in which the first area ratio exceeds a threshold value set based on the second area ratio.

[0014] Preferably, the first feature quantity, the second feature quantity and the third feature quantity have the same dimension and can be compared in the same feature quantity space. The third feature quantity information is the information of the representative positions of multiple third feature quantities in the feature quantity space. The processor performs the following processing: calculates the first distance between the representative position in the feature quantity space and the position of the first feature quantity, calculates the third distance between the representative positions of multiple second feature quantities in the feature quantity space and the position of the first feature quantity, and makes a judgment based on the first distance and the third distance.

[0015] Preferably, the processor determines that an abnormality exists in the tissue slice shown in the first slice image when the weighted sum of the first distance and the third distance is outside a set range.

[0016] Preferably, the processor performs the following processing: receiving an instruction to set a weight coefficient to be applied to the first distance and the third distance in the weighted sum.

[0017] Preferably, the first slice image and the second slice image are images obtained by photographing a slide specimen carrying tissue slices of multiple organs, and the processor performs the following processing: identifying the tissue slices of each organ from the first slice image and the second slice image, extracting the first feature value and the second feature value of each of the identified tissue slices of each organ, and determining whether there is an abnormality in each of the tissue slices of each organ reflected in the first slice image based on the first feature value, the second feature value and the third feature value information.

[0018] The working method of the drug development support device of the present invention includes the following steps: acquiring a first slice image, the first slice image showing a tissue slice of at least one subject among a plurality of subjects constituting a drug administration group of candidate substances to which drugs are administered; extracting a first feature quantity from the first slice image; acquiring a plurality of second slice images, the plurality of second slice images showing tissue slices of a plurality of subjects, the plurality of subjects having the same attributes as the subjects constituting the drug administration group and being in the same breeding environment, and constituting a control group of candidate substances not administered with drugs; extracting a second feature quantity from each of the plurality of second slice images; acquiring a third feature quantity information related to a third feature quantity, the third feature quantity being extracted from each of a plurality of third slice images showing tissue slices of a plurality of subjects constituting a control group not administered with candidate substances in past evaluation tests; and determining whether there is an abnormality in the tissue slice shown in the first slice image based on the first feature quantity, the second feature quantity and the third feature quantity information.

[0019] The working program of the drug development support device of the present invention causes a computer to execute a process including the following steps: acquiring a first slice image, the first slice image showing a tissue slice of at least one subject among a plurality of subjects constituting a drug administration group of a candidate substance to be administered with a drug; extracting a first feature quantity from the first slice image; acquiring a plurality of second slice images, the plurality of second slice images showing tissue slices of a plurality of subjects, the plurality of subjects having the same attributes as the subjects constituting the drug administration group and being in the same breeding environment, and constituting a control group of the candidate substance not administered with the drug; extracting a second feature quantity from each of the plurality of second slice images; acquiring a third feature quantity information related to a third feature quantity, the third feature quantity being extracted from each of a plurality of third slice images showing tissue slices of a plurality of subjects constituting a control group to which the candidate substance was not administered in a past evaluation test; and determining whether there is an abnormality in the tissue slice shown in the first slice image based on the first feature quantity, the second feature quantity and the third feature quantity information.

[0020] Effects of the Invention

[0021] According to the technology of the present invention, it is possible to provide a drug development support device, a method for operating the drug development support device, and a program for operating the drug development support device that can determine whether an abnormality caused by a candidate substance of a drug occurs in a subject in a drug administration group. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 The diagram shows the process of an evaluation test, slice images, and a drug development support device.

[0023] Figure 2 It is a graph showing the current drug-administered group and the current control group.

[0024] Figure 3 This is a block diagram showing a computer constituting the drug development support device.

[0025] Figure 4 This is a block diagram showing a processing unit of a CPU of a drug development support device.

[0026] Figure 5 A diagram showing a first block image obtained by subdividing the first slice image.

[0027] Figure 6 This is a diagram showing a state where the first feature quantity is extracted from the first block image by the feature quantity extractor.

[0028] Figure 7 This is a diagram showing a state where the second feature quantity is extracted from the second block image by the feature quantity extractor.

[0029] Figure 8 This is a diagram showing the structure of a feature extractor.

[0030] Fig. 9 is a diagram showing the processing in the learning phase of the autoencoder.

[0031] Fig.10 is a graph showing the past control group.

[0032] Fig.11 This is a diagram showing a state where the third feature quantity is extracted from the third block image by the feature quantity extractor.

[0033] Fig.12 This is a diagram showing a graph in which the third feature quantity is plotted in the feature quantity space and the third feature quantity information.

[0034] Fig.13 It is a diagram showing the first distance and the second distance.

[0035] Fig.14 This is a diagram showing a state in which the comparison result between the first distance and the set range is outputted in the determination unit.

[0036] Fig.15 This is a diagram showing a state in which the comparison result between the second distance and the set range is outputted in the determination unit.

[0037] Fig.16 This is a diagram showing the first area ratio.

[0038] Fig.17 It is a diagram showing a state where a plurality of second area ratios are calculated based on a plurality of second slice images.

[0039] Fig.18 is a diagram showing the processing of the determination unit, Fig.18A indicates that the p-value in the statistical hypothesis test using the first area ratio and the second area ratio is greater than the significance level. Fig.18 B indicates that the p-value is less than the significance level.

[0040] Fig.19 This is a diagram showing a drug development support screen before an abnormality determination instruction is issued.

[0041] Fig. 20 This is a diagram showing a drug development support screen after an abnormality determination instruction is issued.

[0042] Fig.21 This is a flowchart showing the processing procedure of the drug development support device.

[0043] Fig. 22 This is a diagram showing another example of the processing of the determination unit.

[0044] Fig.23 This is a diagram showing the processing of the determination unit in the 2_1st embodiment.

[0045] Fig.24 It is a diagram showing the 2_2 implementation mode of the receiving coefficient setting indication.

[0046] Fig.25 It is a diagram showing a third embodiment of processing slice images obtained by imaging a slide specimen on which tissue slices of various organs are mounted.

[0047] Fig.26 This is a diagram showing a state where the first feature quantity is extracted from the first block image obtained by subdividing the cardiac slice by the feature quantity extractor for the cardiac slice.

[0048] Fig. 27 This is a diagram showing another example of the drug development support screen. DETAILED DESCRIPTION

[0049] [First embodiment]

[0050] As an example, Figure 1 As shown, the drug development support device 10 of the present invention is used to evaluate candidate substances 26 of a drug (refer to Figure 2 )'s efficacy and toxicity. The drug development support device 10 is, for example, a desktop personal computer, and includes a display 11 for displaying various screens and an input device 12 composed of a keyboard, a mouse, a touch panel and / or a microphone for voice input. The drug development support device 10 is operated by an operator such as a staff member of a pharmaceutical facility involved in drug development.

[0051] The slice image 15 is input to the drug development support device 10. The slice image 15 is generated, for example, by the following sequence. First, a subject S such as a rat prepared for evaluating a candidate substance 26 is dissected, and organs of the subject S are collected, and multiple tissue slices (hereinafter referred to as liver slices) LVS of the cross section of the liver LV are collected. Next, after the collected liver slices LVS are pasted one by one on a slide 16, the liver slices LVS are stained, and in this case, the liver slices LVS are stained with hematoxylin-eosin. Next, the stained liver slices LVS are covered with a cover glass 17 to complete a slide specimen 18. Then, the slide specimen 18 is placed on a photographic device 19 such as a digital optical microscope, and the photographic device 19 photographs the slice image 15. In the slice image 15 obtained in this way, a subject ID (Identification Data) for uniquely identifying the subject S, an image ID for uniquely identifying the slice image 15, and a photographing date, etc. are added. In addition, the staining may be staining with hematoxylin pigment alone, staining with nuclear fast red pigment, etc.

[0052] As an example, Figure 2 As shown, the slice images 15 input to the drug development support device 10 include a first slice image 151 and a second slice image 152. The first slice image 151 is an image obtained by photographing a liver slice LVS of a subject S of the current dosing group 25. The current dosing group 25 is composed of a plurality of subjects S to which a candidate substance 26 is administered. The number of subjects S constituting the current dosing group 25 is, for example, about 5 to 10. A plurality of first slice images 151 are obtained from one subject S, and therefore, the number of first slice images 151 obtained by multiplying the number obtained from one subject S by the number of subjects S is obtained from the current dosing group 25. For example, when the number obtained from one subject S is 10 and the number of subjects S constituting the current dosing group 25 is 10, 10×10=100 first slice images 151 can be obtained.

[0053] The second slice image 152 is an image obtained by photographing a liver slice LVS of the subject S of the current control group 27. In contrast to the current medication group 25, the current control group 27 is composed of a plurality of subjects S to which the candidate substance 26 is not administered. The subjects S constituting the current control group 27 have the same attributes as the subjects S constituting the current medication group 25, and are subjects S in the same breeding environment. The same attributes refer to, for example, the same age and / or the same sex. In addition, the same attributes also include the same age composition ratio and / or the same sex composition ratio (5 males and 5 females, etc.). The same breeding environment refers to, for example, the same feed, the same temperature and humidity of the breeding space, and / or the same width of the breeding space. The "same" of the same breeding environment refers to the same meaning of the error that is generally allowed in the technical field to which the technology of the present invention belongs and does not deviate from the technical spirit of the present invention, in addition to being completely the same. The number of subjects S constituting the current control group 27 is also the same as that of the current medication group 25, for example, about 5 to 10. Similar to the first slice images 151 , a plurality of second slice images 152 are obtained from one subject S. Therefore, a number of second slice images 152 obtained by multiplying the number obtained from one subject S by the number of subjects S is obtained from the current control group 27 .

[0054] In addition, in the current administration group 25, there are a plurality of groups with different administration amounts of the candidate substance 26. For example, the administration amount of the candidate substance 26 is made different in three stages, namely, a high administration amount group, a medium administration amount group, and a low administration amount group. In this way, the effect of the administration amount of the candidate substance 26 on the subject S can be determined.

[0055] As an example, Figure 3 As shown, the computer constituting the drug development support apparatus 10 includes a storage device 30 , a memory 31 , a CPU (Central Processing Unit) 32 , and a communication unit 33 in addition to the display 11 and the input device 12 . These are connected to each other via a bus 34 .

[0056] The storage device 30 is a hard disk drive built into the computer constituting the drug development support device 10 or connected via a cable or a network. Alternatively, the storage device 30 is a disk array formed by connecting multiple hard disk drives. The storage device 30 stores control programs such as an operating system, various application programs, and various data attached to these programs. In addition, a solid-state drive can also be used instead of a hard disk drive.

[0057] The memory 31 is a working memory for the CPU 32 to perform processing. The CPU 32 loads the program stored in the storage device 30 into the memory 31, and performs processing according to the program. Thus, the CPU 32 centrally controls each part of the computer. The CPU 32 is an example of a "processor" involved in the technology of the present invention. In addition, the memory 31 can be built into the CPU 32. The communication unit 33 performs transmission control of various information with external devices such as the photographic device 19.

[0058] As an example, Figure 4 As shown, a working program 40 is stored in the storage device 30 of the drug development support device 10. The working program 40 is an application program for causing the computer to function as the drug development support device 10. That is, the working program 40 is an example of a "working program of a drug development support device" involved in the technology of the present invention. A feature quantity extractor 41, third feature quantity information 42, and setting range information 43 are also stored in the storage device 30. The feature quantity extractor 41 is an example of a "machine learning model" involved in the technology of the present invention.

[0059] If the work program 40 is started, the CPU 32 of the computer constituting the drug development support device 10 cooperates with the memory 31, etc. to function as a read-write (hereinafter referred to as RW (Read Write)) control unit 50, a first feature value extraction unit 511, a second feature value extraction unit 512, a judgment unit 52 and a display control unit 53.

[0060] The RW control unit 50 controls storage of various data in the storage device 30 and reading of various data in the storage device 30. For example, the RW control unit 50 stores the first slice image 151 and the second slice image 152 from the imaging device 19 in the storage device 30.

[0061] The RW control unit 50 obtains the first slice image 151 by reading it from the storage device 30, and outputs the first slice image 151 to the first feature extraction unit 511 and the display control unit 53. The first slice image 151 output from the RW control unit 50 to the first feature extraction unit 511 is a target for determining whether the liver slice LVS has an abnormality. Hereinafter, the first slice image 151, which is a target for determining whether the liver slice LVS has an abnormality, is denoted as the target first slice image 151T (reference number 151T). Figure 5 wait).

[0062] The RW control unit 50 acquires the plurality of second slice images 152 by reading them from the storage device 30 . The RW control unit 50 outputs all the second slice images 152 to the second feature quantity extraction unit 512 .

[0063] The RW control unit 50 reads the feature quantity extractor 41 from the storage device 30, and outputs the feature quantity extractor 41 to the first feature quantity extractor 511 and the second feature quantity extractor 512. In addition, the RW control unit 50 obtains the third feature quantity information 42 by reading it from the storage device 30, and outputs the third feature quantity information 42 to the determination unit 52. In addition, the RW control unit 50 reads the setting range information 43 from the storage device 30, and outputs the setting range information 43 to the determination unit 52.

[0064] The first feature quantity extraction unit 511 extracts the first feature quantity 601 from the target first slice image 151T through the feature quantity extractor 41. The first feature quantity extraction unit 511 outputs the first feature quantity 601 to the determination unit 52.

[0065] The second feature amount extracting unit 512 extracts the second feature amount 602 from each of the second slice images 152 by using the feature amount extractor 41. The second feature amount extracting unit 512 outputs the second feature amount 602 to the determining unit 52.

[0066] The determination unit 52 determines whether or not there is an abnormality caused by the candidate substance 26 in the liver slice LVS shown in the target first slice image 151T based on the first feature quantity 601, the second feature quantity 602, the third feature quantity information 42, and the set range information 43. The abnormality refers to a lesion not observed in a normal liver slice LVS, such as a tumor, inflammation, cyst, and infiltration. The determination unit 52 outputs a determination result 61 of whether or not there is an abnormality caused by the candidate substance 26 in the liver slice LVS shown in the target first slice image 151T to the display control unit 53.

[0067] The display control unit 53 controls the display of various screens on the display 11. Among the various screens, there is a drug development support screen 100 (see FIG. 100 ) that displays the first slice image 151 and the determination result 61. Fig.19 and Fig. 20 ) etc. In addition, in the CPU 32, in addition to the processing units 50 to 53, an instruction receiving unit 110 (see Fig.24 )wait.

[0068] As an example, Figure 5As shown in FIG. 1 , the first feature quantity extraction unit 511 uses a well-known image recognition technology to recognize the liver slice LVS shown in the target first slice image 151T, and subdivides the recognized liver slice LVS into a plurality of first block images 651. The first block image 651 has a predetermined size that can be processed by the feature quantity extractor 41. Although not shown in the figure, the second feature quantity extraction unit 512 similarly subdivides the liver slice LVS shown in the second slice image 152 into a plurality of second block images 652 (see FIG. 1 ). Figure 7 ). In addition, Figure 5 In the example, the first block image 651 does not have an area overlapping with other first block images 651, but the first block image 651 may overlap with a part of other first block images 651. The second block image 652 is the same.

[0069] As an example, Figure 6 As shown, the first feature quantity extraction unit 511 extracts the first feature quantity 601 from each of the plurality of first block images 651 obtained by subdividing the target first slice image 151T through the feature quantity extractor 41. Figure 7 As shown, the second feature amount extracting unit 512 extracts the second feature amount 602 for each of the plurality of second block images 652 obtained by segmenting all the second slice images 152 using the feature amount extractor 41 .

[0070] As an example, Figure 8 As shown in FIG. 4 , the encoder unit 71 of the auto encoder 70 is used in the feature quantity extractor 41. The auto encoder 70 includes a decoder unit 72 in addition to the encoder unit 71. Block images 65 such as the first block image 651 and the second block image 652 are input to the encoder unit 71. The encoder unit 71 converts the block image 65 into the feature quantity 60. The encoder unit 71 transmits the feature quantity 60 to the decoder unit 72. The decoder unit 72 generates a restored image 73 of the block image 65 based on the feature quantity 60.

[0071] As is well known, the encoder unit 71 includes a convolution layer for performing convolution processing using a filter and a pooling layer for performing pooling processing such as maximum pooling processing. The same is true for the decoder unit 72. The encoder unit 71 extracts the feature amount 60 by repeatedly performing convolution processing by the convolution layer and pooling processing by the pooling layer on the input block image 65 a plurality of times. The extracted feature amount 60 represents the features of the shape and texture of the liver slice LVS reflected in the block image 65.

[0072] The feature quantity 60 is a set of multiple numerical values. That is, the feature quantity 60 is multidimensional data. The dimension of the feature quantity 60 is, for example, 512, 1024, or 2048. The first feature quantity 601, the second feature quantity 602, and the third feature quantity 603 described later (see Fig.11) have the same dimension and can be in the same feature space 80 (reference Fig.12 etc.) for comparison.

[0073] As an example, Fig. 9 As shown, the auto encoder 70 inputs the third block image 653L for learning in the learning stage before the encoder unit 71 is transferred to the feature quantity extractor 41 for learning. The auto encoder 70 outputs a restored image 73L for learning with respect to the third block image 653L for learning. Based on the third block image 653L for learning and the restored image 73L for learning, the loss calculation of the auto encoder 70 using the loss function is performed. Then, based on the result of the loss calculation, the update setting of various coefficients of the auto encoder 70 (coefficients of filters of the convolutional layer, etc.) is performed, and the auto encoder 70 is updated according to the updated setting.

[0074] In the learning phase of the auto encoder 70, the input of the third block image 653L for learning to the auto encoder 70, the output of the restored image 73L for learning from the auto encoder 70, the loss calculation, the update setting, and the update of the auto encoder 70 are repeated while the third block image 653L for learning is replaced. The repetition of the above series of processes ends when the restoration accuracy from the third block image 653L for learning to the restored image 73L for learning reaches a preset level. In this way, the encoder unit 71 of the auto encoder 70 whose restoration accuracy reaches the set level is stored as the feature quantity extractor 41 in the storage device 30 of the drug development support apparatus 10. In addition, the learning may end when the above series of processes are repeated a set number of times regardless of the restoration accuracy from the third block image 653L for learning to the restored image 73L for learning.

[0075] Such learning of the autoencoder 70 may be performed by the drug development support apparatus 10 or by a device different from the drug development support apparatus 10. In the latter case, the feature quantity extractor 41 is sent from another device to the drug development support apparatus 10, and the feature quantity extractor 41 is stored in the storage device 30 by the RW control unit 50.

[0076] As an example, Fig.10As shown, the third block image 653L for learning is provided by a plurality of third block images 653 obtained by subdividing the liver slice LVS shown in the third slice image 153. The third slice image 153 is an image obtained by photographing the liver slice LVS of the subject S of the past control group 75. The past control group 75 is composed of a plurality of subjects S to which the candidate substance of the drug was not administered in the past evaluation test. Therefore, the number of subjects S constituting the past control group 75 is much larger than the number of subjects S constituting the current medication group 25 and the current control group 27, for example, about several hundred to several thousand. The third slice image 153 is obtained from a plurality of subjects S in the same manner as the first slice image 151, etc. Therefore, the number of third slice images 153 obtained by multiplying the number obtained from one subject S by the number of subjects S is obtained from the past control group 75.

[0077] Next, the structure of the third feature information 42 is described. First, as an example, Fig.11 As shown, a plurality of third feature quantities 603 are respectively extracted by the feature quantity extractor 41 from a plurality of third block images 653 based on a plurality of third slice images 153 obtained from all the subjects S constituting the past control group 75 .

[0078] Fig.12 is drawn in the feature space 80 Fig.11 The third feature quantity information 42 includes coordinates 81 in the feature quantity space 80 of the representative position of the third feature quantity 603 indicated by the × mark (hereinafter referred to as representative position coordinates). The representative position is, for example, the center point or average point of the distribution 82 of the third feature quantity 603. Fig.12 In the figure, for the sake of convenience, the dimension of the feature space 80 is set to be two-dimensional with a D1 axis and a D2 axis, but the actual dimension of the feature space 80 is 512 dimensions, etc. as described above. Fig.13 Likewise, for convenience of explanation, the dimension of the feature space 80 is expressed in two dimensions.

[0079] Similar to the learning of the autoencoder 70, the third feature quantity information 42 may be produced by the drug development support device 10 or by a device different from the drug development support device 10. In the latter case, the third feature quantity information 42 is sent from another device to the drug development support device 10, and the third feature quantity information 42 is stored in the storage device 30 by the RW control unit 50. In addition, the third feature quantity information 42 may be a plurality of third feature quantities 603 themselves. In this case, the representative position coordinates 81 are derived by the drug development support device 10.

[0080] As an example, Fig.13As shown, the double-dashed line frame and the symbol 85 indicate the distribution of the first feature quantity 601, which is extracted for each of the plurality of first block images 651 obtained by subdividing the first target slice image 151T. Also, the double-dashed line frame and the symbol 86 indicate the distribution of the second feature quantity 602, which is extracted for each of the plurality of second block images 652 obtained by subdividing all the second slice images 152.

[0081] The determination unit 52 calculates a first distance D1 in the feature quantity space 80 between the representative position of the third feature quantity 603 represented by the representative position coordinate 81 of the third feature quantity information 42 and the position of the first feature quantity 601. The determination unit 52 calculates the first distance D1 of a plurality of first feature quantities 601 extracted for each of a plurality of first block images 651 obtained by subdividing the first slice image 151T of the target. Furthermore, the determination unit 52 calculates a second distance D2 in the feature quantity space 80 between the representative position of the third feature quantity 603 and the position of the second feature quantity 602. The determination unit 52 calculates a second distance D2 of a plurality of second feature quantities 602 extracted for each of a plurality of second block images 652 obtained by subdividing all the second slice images 152. The first distance D1 and the second distance D2 are Mahalanobis distances.

[0082] As an example, Fig.14 As shown, the determination unit 52 compares the calculated first distance D1 with the setting range of the setting range information 43. The setting range information 43 includes a threshold value indicating the setting range. When the first distance D1 is less than the threshold value of the setting range information 43, the determination unit 52 outputs a comparison result 90 indicating that the first distance D1 is within the setting range. On the other hand, when the first distance D1 is greater than the threshold value of the setting range information 43, the determination unit 52 outputs a comparison result 90 indicating that the first distance D1 is outside the setting range.

[0083] As an example, Fig.15 As shown, the determination unit 52 compares the calculated second distance D2 with the setting range of the setting range information 43. When the second distance D2 is less than the threshold value of the setting range information 43, the determination unit 52 outputs a comparison result 90 indicating that the second distance D2 is within the setting range. On the other hand, when the second distance D2 is greater than the threshold value of the setting range information 43, the determination unit 52 outputs a comparison result 90 indicating that the second distance D2 is outside the setting range.

[0084] As an example, Fig.16As shown in FIG. 1 , the determination unit 52 calculates a first area ratio 951 of the first block images 651 (indicated by hatching) whose first distance D1 is outside the set range in the target first slice image 151T. Specifically, the determination unit 52 calculates the first area ratio 951 by dividing the number of the first block images 651 whose first distance D1 is outside the set range by the total number of the first block images 651. Fig.16 , the case where the first area ratio 951 is 11.9% is illustrated.

[0085] As in the case of the first area ratio 951, the determination unit 52 calculates the second area ratio 952 of the second block image 652 whose second distance D2 is outside the set range in the second slice image 152. However, as described above, the second distance D2 is calculated based on all the second slice images 152, so the determination unit 52 calculates the second area ratio 952 as an example. Fig.17 As shown, the second area ratio 952 is calculated for each of the second slice images 152 .

[0086] As an example, Fig.18 As shown, the determination unit 52 determines whether there is an abnormality caused by the candidate substance 26 in the liver slice LVS shown in the target first slice image 151T based on the p value in the statistical hypothesis test based on the null hypothesis that there is no difference between the first area ratio 951 and the second area ratio 952. Fig.18 As shown in A, when the p value is greater than the significant level (0.05 in this example), the determination unit 52 regards that there is no statistically significant difference and outputs a determination result 61 indicating that there is no abnormality caused by the candidate substance 26 to the liver slice LVS shown in the target first slice image 151T. On the other hand, Fig.18 As shown in B, when the p value is smaller than the significance level, the determination unit 52 regards that there is a statistically significant difference and outputs a determination result 61 indicating the presence of an abnormality caused by the candidate substance 26 to the liver slice LVS shown in the target first slice image 151T.

[0087] As an example, Fig.19 As shown, the display control unit 53 controls the display of the drug development support screen 100 on the display 11. The display control unit 53 displays the drug development support screen 100 when the operator issues a display instruction of the first slice image 151 through the input device 12 and the display instruction is received by the instruction receiving unit 110. A plurality of first slice images 151 are displayed side by side on the drug development support screen 100. These plurality of first slice images 151 are first slice images 151 obtained from one of the plurality of subjects S constituting the current medication group 25. Fig.19, an example is shown in which 10 first slice images 151 with image IDs "SP001" to "SP010" obtained from a subject S with a subject ID "R001" are displayed side by side. In addition, the subject ID becomes a pull-down menu 101, and the subject S whose first slice image 151 is displayed on the drug development support screen 100 can be switched. In addition, the pull-down menu 101 has an option for specifying all of the multiple subjects S constituting the current medication group 25.

[0088] An abnormality determination button 102 is provided at the bottom of the drug development support screen 100. When the abnormality determination button 102 is selected by the operator, an abnormality determination instruction is received by the instruction receiving unit 110. The abnormality determination instruction is an instruction for causing the first feature quantity extraction unit 511, the second feature quantity extraction unit 512, and the determination unit 52 to extract the first feature quantity 601 and the second feature quantity 602, and to determine whether there is an abnormality in the liver slice LVS shown in the first slice image 151 displayed on the drug development support screen 100. That is, the first slice image 151 displayed on the drug development support screen 100 corresponds to the target first slice image 151T.

[0089] When the abnormality determination button 102 is selected and the determination unit 52 determines whether or not an abnormality exists in the liver slice LVS shown in the target first slice image 151T, as an example, Fig. 20 As shown in FIG. 1 , the display control unit 53 displays the determination result 61 at the bottom of each first slice image 151. Fig. 20 , a case where abnormality is determined to exist in the liver slice LVS shown in three first slice images 151 of image IDs “SP007” to “SP009” is illustrated. In addition, a frame of the first block image 651 indicating that the comparison result 90 is outside the set range may be superimposed on the first slice image 151 and displayed.

[0090] Next, as an example, refer to Fig.21 The flowchart shown in FIG. 1 illustrates the function of the above structure. First, if the working program 40 is started in the drug development support device 10, as shown in FIG. Figure 4 As shown, the CPU 32 of the drug development support device 10 functions as an RW control unit 50 , a first feature value extraction unit 511 , a second feature value extraction unit 512 , a determination unit 52 , and a display control unit 53 .

[0091] First, the first slice image 151 and the second slice image 152 are captured by the imaging device 19. The first slice image 151 is obtained by capturing the liver slice LVS of the subject S in the current medication group 25, and the second slice image 152 is obtained by capturing the liver slice LVS of the subject S in the current control group 27. The first slice image 151 and the second slice image 152 are output from the imaging device 19 to the drug development support device 10. In the drug development support device 10, the RW control unit 50 stores the first slice image 151 and the second slice image 152 from the imaging device 19 in the storage device 30 (step ST100).

[0092] When the operator issues a display instruction for the first slice image 151 via the input device 12 ("Yes" in step ST110), the first slice image 151 designated by the display instruction is read from the storage device 30 via the RW control unit 50 (in step ST120). The first slice image 151 is output from the RW control unit 50 to the display control unit 53. Fig.19 As shown, the first slice image 151 is displayed on the display 11 through the drug development support screen 100 under the control of the display control unit 53 (step ST130 ).

[0093] When the operator selects the abnormality determination button 102 of the drug development support screen 100 and issues an abnormality determination instruction ("Yes" in step ST140), the RW control unit 50 reads the first slice image 151 displayed on the drug development support screen 100 as the target first slice image 151T from the storage device 30 to obtain (step ST150). In addition, the RW control unit 50 reads all the second slice images 152 from the storage device 30 to obtain (step ST150). The target first slice image 151T is output from the RW control unit 50 to the first feature quantity extraction unit 511. In addition, the second slice image 152 is output from the RW control unit 50 to the second feature quantity extraction unit 512.

[0094] The feature quantity extractor 41 is read from the storage device 30 by the RW control unit 50 (step ST150), and is output to the first feature quantity extractor 511 and the second feature quantity extractor 512. Furthermore, the third feature quantity information 42 is obtained by reading it from the storage device 30 by the RW control unit 50 (step ST150), and is output to the determination unit 52. Furthermore, the setting range information 43 is read from the storage device 30 by the RW control unit 50 (step ST150), and is output to the determination unit 52.

[0095] like Figure 5As shown, in the first feature quantity extraction unit 511, the first slice image 151 is subdivided into a plurality of first block images 651 (step ST160). Similarly, in the second feature quantity extraction unit 512, the second slice image 152 is subdivided into a plurality of second block images 652 (step ST160).

[0096] Then, if Figure 6 As shown, in the first feature quantity extraction unit 511, the first feature quantity 601 is extracted from the first block image 651 by the feature quantity extractor 41 (step ST170). The first feature quantity 601 is output from the first feature quantity extraction unit 511 to the determination unit 52. Figure 7 As shown, in the second feature quantity extraction unit 512, the second feature quantity 602 is extracted from the second block image 652 by the feature quantity extractor 41 (step ST170). The second feature quantity 602 is output from the second feature quantity extraction unit 512 to the determination unit 52.

[0097] like Fig.13 As shown, in the determination unit 52, the first distance D1 between the representative position of the third feature quantity 603 and the position of the first feature quantity 601 is calculated (step ST180). In addition, the second distance D2 between the representative position of the third feature quantity 603 and the position of the second feature quantity 602 is calculated (step ST180).

[0098] like Fig.14 As shown, in the determination unit 52, the first distance D1 is compared with the setting range of the setting range information 43, and a comparison result 90 is output. Fig.15 As shown, the second distance D2 is compared with the setting range of the setting range information 43, and a comparison result 90 is output. Fig.16 As shown in FIG. 1 , the first area ratio 951 of the first block image 651 whose first distance D1 is outside the set range in the first slice image 151 is calculated (step ST190). Fig.17 As shown, a second area ratio 952 of the second block image 652 whose second distance D2 is outside the set range in the second slice image 152 is calculated (step ST190 ).

[0099] like Fig.18 As shown in FIG. 1 , the determination unit 52 determines whether or not there is an abnormality caused by the candidate substance 26 in the liver slice LVS shown in the target first slice image 151T based on the p value in the statistical hypothesis test using the first area ratio 951 and the second area ratio 952 (step ST200). The determination result 61 is output from the determination unit 52 to the display control unit 53. Fig. 20As shown, the determination result 61 is displayed on the display 11 through the drug development support screen 100 under the control of the display control unit 53 (step ST210). The operator reads the determination result 61, observes in detail the first slice image 151 determined to be abnormal, and evaluates the candidate substance 26 based on the determination result 61.

[0100] As described above, the CPU 32 of the drug development support device 10 includes the RW control unit 50, the first feature extraction unit 511, the second feature extraction unit 512, and the determination unit 52. The RW control unit 50 acquires the first slice image 151, the second slice image 152, and the third feature information 42 by reading from the storage device 30. The first slice image 151 is an image obtained by photographing the liver slice LVS of the subject S constituting the current administration group 25 of the candidate substance 26 to which the drug is administered. The second slice image 152 is a plurality of images obtained by photographing the liver slice LVS of a plurality of subjects S, the plurality of subjects S having the same attributes as the subject S constituting the current administration group 25 and being in the same breeding environment, and constituting the current control group 27 of the candidate substance 26 not administered with the drug. The third feature information 42 is information on the third feature 603 extracted from the third slice image 153. The third slice image 153 is a plurality of images obtained by imaging liver slices LVS of a plurality of subjects S constituting a past control group 75 to which the candidate substance was not administered in a past evaluation test.

[0101] The first feature quantity extraction unit 511 extracts the first feature quantity 601 from the target first slice image 151T. The second feature quantity extraction unit 512 extracts the second feature quantity 602 from each of the plurality of second slice images 152. The determination unit 52 determines whether there is an abnormality in the liver slice LVS shown in the target first slice image 151T based on the first feature quantity 601, the second feature quantity 602, and the third feature quantity information 42.

[0102] The current control group 27 includes subjects S that have abnormalities due to properties and / or the breeding environment. The same is true for the past control group 75. In the technology of the present invention, by comparing the second feature quantity 602 and the third feature quantity information 42 based on these current control group 27 and the past control group 75 with the first feature quantity 601 based on the current medication group 25, it is determined whether there is an abnormality in the liver slice LVS of the subject S constituting the current medication group 25. Therefore, in the case where the determination result 61 output by such a determination has an abnormality, it can be known that the cause of the abnormality is not the properties and / or the breeding environment, in other words, the cause of the abnormality is the candidate substance 26. Therefore, according to the technology of the present invention, it is possible to determine whether the abnormality caused by the candidate substance 26 of the drug occurs in the subject S of the current medication group 25.

[0103] The subjects S constituting the current control group 27 have the same properties and are in the same breeding environment as the subjects S constituting the current medication group 25, but because of the small number, it is unreliable to make a judgment based on this alone. Therefore, in the technology of the present invention, the past control group 75 is used to make up for the defect of the small number of subjects S constituting the current control group 27. On the other hand, the number of subjects S constituting the current medication group 25 is large, but sometimes the properties and breeding environment of the subjects S constituting the past control group 75 are different. Therefore, if the judgment is made based on this alone, it is possible that an increase in false judgments will occur. Therefore, in the technology of the present invention, the current control group 27 is used to make up for the differences in the properties and breeding environment of the subjects S constituting the past control group 75. That is, the current control group 27 and the past control group 75 are in a complementary relationship. In the technology of the present invention, by appropriately combining the current control group 27 and the past control group 75 that are in such a complementary relationship, it is possible to determine whether the cause of the abnormality occurring in the subject S is the candidate substance 26.

[0104] The first feature quantity extraction unit 511 and the second feature quantity extraction unit 512 extract the first feature quantity 601 and the second feature quantity 602 by using the feature quantity extractor 41 of the encoder unit 71 of the auto encoder 70 that has been learned using a plurality of third slice images 153. Therefore, the first feature quantity 601 and the second feature quantity 602 that well express the features of the shape and texture of the liver slice LVS can be easily extracted.

[0105] The first feature quantity 601, the second feature quantity 602, and the third feature quantity 603 have the same dimension and can be compared in the same feature quantity space 80. The third feature quantity information 42 is information on representative positions of the plurality of third feature quantities 603 in the feature quantity space 80. The determination unit 52 calculates a first distance D1 between the representative position of the third feature quantity 603 in the feature quantity space 80 and the position of the first feature quantity 601. Furthermore, the determination unit 52 calculates a second distance D2 between the representative position of the third feature quantity 603 in the feature quantity space 80 and the respective positions of the plurality of second feature quantities 602. The determination unit 52 makes a determination based on the first distance D1 and the second distance D2.

[0106] The first distance D1 indicates the degree of deviation between the first slice image 151T and the third slice image 153 of the object, and the second distance D2 indicates the degree of deviation between the second slice image 152 and the third slice image 153. The degree of deviation between the first slice image 151T and the third slice image 153 of the object indicates the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the past control group 75. Furthermore, the degree of deviation between the second slice image 152 and the third slice image 153 indicates the degree of difference between the abnormality occurring in the current control group 27 and the abnormality occurring in the past control group 75. Therefore, if the judgment is made based on the first distance D1 and the second distance D2, the judgment can be made taking into account both the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the past control group 75 and the degree of difference between the abnormality occurring in the current control group 27 and the abnormality occurring in the past control group 75, and the judgment accuracy can be improved.

[0107] The first feature quantity extraction unit 511 extracts the first feature quantity 601 for each of the plurality of first block images 651 obtained by subdividing the target first slice image 151T. Furthermore, the second feature quantity extraction unit 512 extracts the second feature quantity 602 for each of the plurality of second block images 652 obtained by subdividing the second slice image 152. The determination unit 52 calculates the first distance D1 for each of the plurality of first feature quantities 601 extracted from the plurality of first block images 651. Furthermore, the determination unit 52 calculates the second distance D2 for each of the plurality of second feature quantities 602 extracted from the plurality of second block images 652. The determination unit 52 makes a determination based on the first area ratio 951 of the first block image 651 whose first distance D1 is outside the set range in the target first slice image 151T and the second area ratio 952 of the second block image 652 whose second distance D2 is outside the set range in the second slice image 152. Compared with the case where the determination is performed without segmenting the target first slice image 151T and the second slice image 152 into the first block image 651 and the second block image 652 , the information used as a reference for determination is greatly increased, and thus the determination accuracy can be further improved.

[0108] The greater the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the past control group 75 (the more the abnormality occurring in the current medication group 25 is caused by the candidate substance 26), the greater the value of the first area ratio 951. On the other hand, the smaller the degree of difference between the abnormality occurring in the current control group 27 and the abnormality occurring in the past control group 75 (the similarity between the abnormality occurring in the current control group 27 and the abnormality occurring in the past control group 75), the smaller the value of the second area ratio 952. Therefore, if it is considered that there is a statistically significant difference between the first area ratio 951 and the second area ratio 952, it can be said that there is a high possibility that the abnormality caused by the candidate substance 26 exists in the liver slice LVS shown in the target first slice image 151T. Therefore, in the case where there is a statistically significant difference between the first area ratio 951 and the second area ratio 952, by determining that there is an abnormality in the liver slice LVS shown in the target first slice image 151T, the determination accuracy can be further improved.

[0109] In addition, as a criterion for determination, as an example, Fig. 22 As shown in FIG. 9 , the threshold value TH may be set based on the second area ratio 952. Fig. 22 , as the threshold value TH, the maximum value of the plurality of second area ratios 952 calculated from the plurality of second slice images 152 is exemplified. In this case, the determination unit 52 determines that there is an abnormality caused by the candidate substance 26 in the liver slice LVS shown in the target first slice image 151T where the first area ratio 951 exceeds the threshold value TH. According to this embodiment, since it is not necessary to derive a p value or compare the p value with a significant level, the determination can be easily completed. In addition, the threshold value TH may be an average value or a mode value of the plurality of second area ratios 952.

[0110] [Implementation Method 2_1]

[0111] As an example, Fig.23 As shown, in the second embodiment, the determination unit 52 calculates the third distance D3 instead of the second distance D2. The third distance D3 is the distance between the representative positions (indicated by the + sign) of the plurality of second feature quantities 602 in the feature quantity space 80 and the position of the first feature quantity 601. The third distance D3 is also the Mahalanobis distance. The representative position of the second feature quantity 602 is Fig.13 The center point or average point of the distribution 86 of the second feature quantity 602 shown.

[0112] The determination unit 52 calculates the weighted sum DS=W1×D1+W3×D3 of the first distance D1 and the third distance D3. Here, W1 is a weight coefficient multiplied by the first distance D1, and W3 is a weight coefficient multiplied by the third distance D3. 0≤W1≤1, 0≤W3≤1, and W1+W3=1. The determination unit 52 compares the weighted sum DS with the setting range of the setting range information 105. The setting range information 105 is stored in the storage device 30 and includes a threshold value indicating the setting range. When the weighted sum DS is less than the threshold value of the setting range, the determination unit 52 outputs a comparison result 106 indicating that the weighted sum DS is within the setting range. On the other hand, when the weighted sum DS is greater than the threshold value of the setting range, the determination unit 52 outputs a comparison result 106 indicating that the weighted sum DS is outside the setting range.

[0113] The determination unit 52 determines the configuration Fig.13 For each of the first feature quantities 601 of the distribution 85 shown, the first distance D1, the third distance D3, and the weighted sum DS are calculated, and the comparison result 106 is output. When even one of the weighted sums DS calculated from the plurality of first feature quantities 601 constituting the distribution 85 indicates a comparison result 106 outside the set range, the determination unit 52 outputs a determination result 61 indicating the presence of an abnormality caused by the candidate substance 26 to the liver slice LVS projected on the target first slice image 151T.

[0114] Thus, in the second first embodiment, the determination unit 52 calculates the first distance D1 between the representative position of the third feature quantity 603 in the feature quantity space 80 and the position of the first feature quantity 601, and calculates the third distance D3 between the representative positions of the plurality of second feature quantities 602 in the feature quantity space 80 and the position of the first feature quantity 601. Then, the determination unit 52 makes a determination based on the first distance D1 and the third distance D3.

[0115] As described above, the first distance D1 indicates the degree of deviation between the first slice image 151T and the third slice image 153 of the object, and the degree of deviation between the first slice image 151T and the third slice image 153 of the object indicates the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the past control group 75. Furthermore, the third distance D3 indicates the degree of deviation between the first slice image 151T and the second slice image 152 of the object, and the degree of deviation between the first slice image 151T and the second slice image 152 of the object indicates the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the current control group 27. Therefore, if the determination is made based on the first distance D1 and the third distance D3, the determination can be made taking into account both the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the past control group 75 and the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the current control group 27, and the determination accuracy can be improved. However, in the second first embodiment, it is necessary to derive the representative position of the second feature quantity 602. Therefore, the processing of the above-mentioned first embodiment is unnecessary and laborious.

[0116] When the weighted sum DS of the first distance D1 and the third distance D3 is outside the set range, the determination unit 52 determines that an abnormality exists in the liver slice LVS shown in the target first slice image 151T. Therefore, it is possible to make a determination taking into account both the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the past control group 75 and the degree of difference between the abnormality occurring in the current medication group 25 and the abnormality occurring in the current control group 27, and the determination accuracy can be improved.

[0117] In addition, the determination may be made based on the simple sum D1+D3 of the first distance D1 and the third distance D3 instead of the weighted sum DS. Furthermore, when the comparison result 106 shows that the area ratio of the first block image 651 outside the set range in the first slice image 151 is greater than the threshold, it may be determined that an abnormality exists in the liver slice LVS shown in the first slice image 151.

[0118] [Implementation Method 2-2]

[0119] As an example, Fig.24As shown, in the 2_2 embodiment, the instruction receiving unit 110 receives a coefficient setting instruction 111 made by an operator through the input device 12. The coefficient setting instruction 111 is an instruction for setting the weight coefficients W1 and W3, including the weight coefficients W1 and W3 set by the operator. When the instruction receiving unit 110 receives the coefficient setting instruction 111, it outputs the coefficient setting information 112 to the determination unit 52. The coefficient setting information 112 is information that directly copies the weight coefficients W1 and W3 included in the coefficient setting instruction 111. The determination unit 52 uses the weight coefficients W1 and W3 of the coefficient setting information 112 to calculate the weighted sum DS. Fig.24 , an example is shown in which 0.2 is set as the weight coefficient W1 and 0.8 is set as the weight coefficient W3.

[0120] When the contents of the current evaluation test involved in the current medication group 25 and the current control group 27 are similar to the contents of the past evaluation test involved in the past control group 75, the operator sets the weight coefficient W1 to a value larger than the weight coefficient W3. On the other hand, when the contents of the current evaluation test involved in the current medication group 25 and the current control group 27 are different from the contents of the past evaluation test involved in the past control group 75, the operator sets the weight coefficient W3 to a value larger than the weight coefficient W1.

[0121] Thus, in the 2-2nd embodiment, the instruction receiving unit 110 receives the coefficient setting instruction 111, which is a setting instruction for the weight coefficients W1 and W3 applied to the first distance D1 and the third distance D3 in the weighted sum DS. Therefore, the operator can set the weight coefficients W1 and W3 adapted to the similarity between the current evaluation test involved in the current medication group 25 and the current control group 27 and the past evaluation test involved in the past control group 75. As a result, the reliability of the determination result 61 can be improved.

[0122] In addition, the similarity between the current evaluation test involving the current medication group 25 and the current control group 27 and the past evaluation test involving the past control group 75 can be derived by comparing the test conditions of the current evaluation test with the test conditions of the past evaluation test, and the weight coefficients W1 and W3 suitable for the derived similarity can be automatically set.

[0123] [Third Embodiment]

[0124] As an example, Fig.25 As shown in FIG. 1 , in the third embodiment, a slide specimen 120 in which tissue sections of various organs are placed on one slide 16 is processed. Fig.25, a case where a heart slice HS, a brain slice BS, and a bone marrow slice BMS are placed in addition to the liver slice LVS is illustrated. In this case, in the slice image 15 (the first slice image 151 and the second slice image 152), in addition to the liver slice LVS, the heart slice HS, the brain slice BS, and the bone marrow slice BMS are also shown.

[0125] The CPU 32 of the drug development support device 10 of the third embodiment functions as a recognition unit 121 in addition to the processing units 50 to 53 of the first embodiment. The recognition unit 121 recognizes the tissue slices of each organ from the slice image 15 (the first slice image 151 and the second slice image 152), for example, using a template or a machine learning model for recognizing the tissue slices of each organ. The recognition unit 121 outputs the coordinate information of the frames 122 to 125 surrounding the tissue slices of each organ as the recognition result. The frame 122 is a frame surrounding the heart slice HS, and the frame 123 is a frame surrounding the liver slice LVS. In addition, the frame 124 is a frame surrounding the brain slice BS, and the frame 125 is a frame surrounding the bone marrow slice BMS.

[0126] Fig.26 The first feature quantity extraction unit 511 divides the heart slice HS in the frame 122 into a plurality of first block images 651, and the feature quantity extractor 130 for the heart slice HS extracts the first feature quantity 601 from the first block image 651. The first feature quantity extraction unit 511 also extracts the first feature quantity 601 for the liver slice LVS in the frame 123, the brain slice BS in the frame 124, and the bone marrow slice BMS in the frame 125. The second feature quantity extraction unit 512 also extracts the second feature quantity 602 in the same manner. That is, the first feature quantity extraction unit 511 and the second feature quantity extraction unit 512 extract the first feature quantity 601 and the second feature quantity 602 for each tissue slice of each organ identified by the identification unit 121. The following processing of the determination unit 52 is the same as the processing shown in the first embodiment or the second-first embodiment, except that the tissue slices of each organ shown in the target first slice image 151T are determined separately, and therefore the illustration and description are omitted.

[0127] Thus, in the third embodiment, the first slice image 151 and the second slice image 152 are images obtained by photographing the slide specimen 120 on which tissue slices of various organs are placed. The recognition unit 121 recognizes the tissue slices of each organ from such first slice image 151 and second slice image 152. The first feature quantity extraction unit 511 and the second feature quantity extraction unit 512 extract the first feature quantity 601 and the second feature quantity 602 of each of the recognized tissue slices of each organ. The determination unit 52 determines whether there is an abnormality in each of the tissue slices of each organ shown in the target first slice image 151T based on the first feature quantity 601, the second feature quantity 602 and the third feature quantity information 42. Therefore, it is also possible to cope with the slide specimen 120 on which tissue slices of various organs are placed. Compared with the slide specimen 18 on which a tissue slice of a single organ is mounted as in the first embodiment, the slide specimen 120 on which tissue slices of multiple organs are mounted as in the third embodiment is more common. Therefore, drug development support more suitable for general use can be performed.

[0128] The frames representing the tissue slices of each organ in the first slice image 151 and the second slice image 152 may be manually delimited by the operator.

[0129] In addition, as an example, Fig. 27 As in the drug development support screen 135 shown in the figure, a check box 136 may be provided on each of the plurality of first slice images 151 displayed, so that the operator can select the first slice image 151 to be the target first slice image 151T. In addition to the abnormality determination button 102, a select all / release button 137 is provided at the bottom of the drug development support screen 135. The select all / release button 137 is a button for selecting all of the plurality of first slice images 151 as the target first slice image 151T, and is also a button for releasing the selection. In this way, the operator can select the target first slice image 151T to be determined. Fig. 27 , a case where only the first slice image 151 of the image ID “SP001” is selected is exemplified. That is, there may be only one first slice image 151T to be determined.

[0130] The encoder unit 71 of the auto encoder 70 is exemplified as the feature quantity extractor 41, but the present invention is not limited thereto. The encoder unit of a convolutional neural network may be used for the feature quantity extractor 41, and the convolutional neural network may learn to classify the input of the block image 65 into a certain category, for example, to output a determination result of whether an abnormality exists in the liver slice LVS shown in the block image 65. In addition, other machine learning models such as a support vector machine may be used for the feature quantity extractor 41.

[0131] In the tissue slice shown in the third learning image 653L, the region that is easily detected as abnormal can be masked and used for learning. Examples of the region that is easily detected as abnormal include the blood vessel region in the liver slice LVS or dust attached when the slide specimen 18 or 120 is prepared.

[0132] The feature amount 60 is not limited to the feature amount extracted by the feature amount extractor 41. It may be an average value, a maximum value, a minimum value, a mode value, or a dispersion of pixel values ​​of the block image 65.

[0133] The organ is not limited to the liver LV and the like as exemplified. It may be the stomach, lung, small intestine or large intestine, etc. Furthermore, the subject S is not limited to rats. It may be mice, guinea pigs, gerbils, hamsters, ferrets, rabbits, dogs, cats or monkeys, etc.

[0134] The technology described in Japanese Patent Application Publication No. 2021-508373 may also be applied. The technology described in Japanese Patent Application Publication No. 2021-508373 determines that an abnormality has occurred in the tissue slice of the subject S of the current administration group 25, but the technology of the present invention shows that when the content in the determination result 61 of the subject S is no abnormality, the cause of the abnormality is not the candidate substance 26, but the properties and the breeding environment.

[0135] like Figure 1 As shown, the drug development support device 10 may be a personal computer installed in a pharmaceutical facility, or may be a server computer installed in a data center independent of the pharmaceutical facility.

[0136] In the case where the drug development support device 10 is constituted by a server computer, the first slice image 151 and the second slice image 152 are sent from a personal computer installed in each pharmaceutical facility to the server computer via a network such as the Internet. The server computer distributes the drug development support screen 100 or 135 to the personal computer in the form of network distribution screen data produced by a markup language such as XML (Extensible Markup Language). The personal computer reproduces the drug development support screen 100 or 135 displayed on the web browser based on the screen data and displays it on the display. In addition, other data description languages ​​such as JSON (Javascript (registered trademark) Object Notation: Javascript object notation) can also be used instead of XML.

[0137] The drug development support device 10 according to the technology of the present invention can be widely used in all stages of drug development, from the initial stage, that is, setting of drug development targets, to the final stage, that is, clinical trials.

[0138] The hardware structure of the computer constituting the drug development support device 10 involved in the technology of the present invention can be modified in various ways. For example, in order to improve processing power and reliability, the drug development support device 10 can also be composed of a plurality of computers separated as hardware. For example, two computers are made to disperse the functions of the first feature extraction unit 511 and the second feature extraction unit 512 and the function of the determination unit 52. In this case, the drug development support device 10 is composed of two computers.

[0139] In this way, the hardware structure of the computer of the drug development support device 10 can be appropriately changed according to the required performance of processing power, safety, reliability, etc. In addition, it is not limited to hardware. For the purpose of ensuring safety and reliability, application programs such as the work program 40 can of course be duplicated or stored in multiple storage devices in a distributed manner.

[0140] In each of the above-mentioned embodiments, for example, as the hardware structure of the processing unit (Processing Unit) that performs various processes such as the RW control unit 50, the first feature extraction unit 511, the second feature extraction unit 512, the determination unit 52, the display control unit 53, the instruction receiving unit 110, and the recognition unit 121, various processors (Processor) shown below can be used. As described above, among the various processors, in addition to a general-purpose processor, i.e., CPU 32, that executes software (work program 40) to function as various processing units, there are also processors such as FPGA (Field Programmable Gate Array) whose circuit structure can be changed after manufacturing, i.e., programmable logic devices (PLD), ASIC (Application Specific Integrated Circuit), etc., which have a circuit structure specially designed for performing specific processing, i.e., dedicated circuits.

[0141] One processing unit may be composed of one of these various processors, or a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs and / or a combination of a CPU and an FPGA). Furthermore, multiple processing units may be composed of one processor.

[0142] As an example of a plurality of processing units constituted by one processor, there is the following method: as represented by computers such as client terminals and servers, one processor is constituted by a combination of one or more CPUs and software, and the processor functions as a plurality of processing units. Second, as represented by a system on chip (SoC), there is the following method: a processor that realizes the functions of the entire system including a plurality of processing units by one IC (Integrated Circuit) chip is used. In this way, various processing units are constituted as a hardware structure using one or more of the above-mentioned various processors.

[0143] Furthermore, as the hardware configuration of these various processors, more specifically, a circuit (circuitry) formed by combining circuit elements such as semiconductor elements can be used.

[0144] The technology of the present invention can also appropriately combine the above-mentioned various embodiments and / or various variations. Moreover, it is not limited to the above-mentioned embodiments, and various structures can of course be adopted as long as it does not deviate from the main purpose. In addition, the technology of the present invention also relates to a storage medium that stores the program non-temporarily in addition to the program.

[0145] The recorded contents and illustrated contents shown above are detailed descriptions of the parts involved in the technology of the present invention, which are only an example of the technology of the present invention. For example, the descriptions related to the above-mentioned structure, function, action and effect are descriptions related to an example of the structure, function, action and effect of the parts involved in the technology of the present invention. Therefore, without departing from the scope of the technical purpose of the present invention, it is of course possible to delete unnecessary parts of the recorded contents and illustrated contents shown above, or to add or replace new elements. In addition, in order to avoid complicated situations and facilitate understanding of the parts involved in the technology of the present invention, the descriptions of technical common sense that does not particularly need to be explained in terms of enabling the implementation of the technology of the present invention are omitted in the recorded contents and illustrated contents shown above.

[0146] In this specification, "A and / or B" has the same meaning as "at least one of A and B". That is, "A and / or B" may mean only A, only B, or a combination of A and B. Furthermore, in this specification, the same idea as "A and / or B" is also applicable to the case of expressing three or more items by connecting them with "and / or".

[0147] All documents, patent applications, and technical standards described in this specification are incorporated herein by reference to the same extent as if each document, patent application, and technical standard was specifically and individually described by reference.

Claims

1. A drug development support device comprising a processor, The processor performs the following processing: acquiring a first slice image showing a tissue slice of at least one subject among a plurality of subjects constituting a drug administration group of a candidate substance to be administered with a pharmaceutical agent; extracting a first feature quantity from the first slice image; Acquiring a plurality of second slice images, wherein the plurality of second slice images represent tissue slices of a plurality of subjects, the plurality of subjects having the same properties as the subjects constituting the administration group and being in the same breeding environment, and constituting a control group of the candidate substance to which no drug is administered; extracting second feature quantities from each of the plurality of second slice images; acquiring third feature quantity information related to a third feature quantity extracted from a plurality of third slice images showing tissue slices of a plurality of subjects constituting a control group to which the candidate substance was not administered in a past evaluation test; and Based on the first feature amount, the second feature amount, and the third feature amount information, it is determined whether or not there is an abnormality in the tissue slice shown in the first slice image.

2. The drug development support device according to claim 1, in, The processor extracts the first feature amount and the second feature amount by using a machine learning model learned using a plurality of the third slice images.

3. The drug development support device according to claim 1 or 2, in, The first feature quantity, the second feature quantity, and the third feature quantity have the same dimension and can be compared in the same feature quantity space. The third feature quantity information is information of representative positions of a plurality of the third feature quantities in the feature quantity space. The processor performs the following processing: calculating a first distance between the representative position and the position of the first feature quantity in the feature quantity space, calculating a second distance between the representative position in the feature quantity space and the positions of each of the plurality of second feature quantities, The determination is performed based on the first distance and the second distance.

4. The drug development support device according to claim 3, in, The processor performs the following processing: extracting the first feature quantity from each of the plurality of first block images obtained by subdividing the first slice image, and calculating the first distance; extracting the second feature quantities from a plurality of second block images obtained by subdividing the second slice image, respectively, and calculating the second distance; and The determination is performed based on a first area ratio of the first block image whose first distance is outside a set range in the first slice image and a second area ratio of the second block image whose second distance is outside the set range in the second slice image.

5. The drug development support device according to claim 4, in, The processor determines that an abnormality exists in the tissue slice shown in the first slice image when there is a statistically significant difference between the first area ratio and the second area ratio.

6. The drug development support device according to claim 4, in, The processor determines that an abnormality exists in the tissue slice that appears in the first slice image in which the first area ratio exceeds a threshold value set based on the second area ratio.

7. The drug development support device according to claim 1 or 2, in, The first feature quantity, the second feature quantity, and the third feature quantity have the same dimension and can be compared in the same feature quantity space. The third feature quantity information is information of representative positions of a plurality of the third feature quantities in the feature quantity space. The processor performs the following processing: calculating a first distance between the representative position and the position of the first feature quantity in the feature quantity space, calculating a third distance between a plurality of representative positions of the second feature quantities in the feature quantity space and a position of the first feature quantity, The determination is performed based on the first distance and the third distance.

8. The drug development support device according to claim 7, in, The processor determines that an abnormality exists in the tissue slice shown in the first slice image when a weighted sum of the first distance and the third distance is outside a set range.

9. The drug development support device according to claim 8, in, The processor receives an instruction to set a weight coefficient to be applied to the first distance and the third distance in the weighted sum.

10. The drug development support device according to any one of claims 1 to 9, in, The first slice image and the second slice image are images obtained by photographing a slide specimen on which tissue slices of various organs are mounted. The processor performs the following processing: identifying tissue slices of each organ from the first slice image and the second slice image, extracting the first feature value and the second feature value of each tissue slice of each identified organ, Based on the first feature amount, the second feature amount, and the third feature amount information, it is determined whether or not there is an abnormality in each of the tissue slices of each organ shown in the first slice image.

11. A method for operating a drug development support device, comprising the following steps: acquiring a first slice image showing a tissue slice of at least one subject among a plurality of subjects constituting a drug administration group of a candidate substance to be administered with a pharmaceutical agent; extracting a first feature quantity from the first slice image; Acquiring a plurality of second slice images, wherein the plurality of second slice images represent tissue slices of a plurality of subjects, the plurality of subjects having the same properties as the subjects constituting the administration group and being in the same breeding environment, and constituting a control group of the candidate substance to which no drug is administered; extracting second feature quantities from each of the plurality of second slice images; acquiring third feature quantity information related to a third feature quantity extracted from a plurality of third slice images showing tissue slices of a plurality of subjects constituting a control group to which the candidate substance was not administered in a past evaluation test; and Based on the first feature amount, the second feature amount, and the third feature amount information, it is determined whether or not there is an abnormality in the tissue slice shown in the first slice image.

12. A working program of a drug development support device, which causes a computer to execute a process comprising the following steps: acquiring a first slice image showing a tissue slice of at least one subject among a plurality of subjects constituting a drug administration group of a candidate substance to be administered with a pharmaceutical agent; extracting a first feature quantity from the first slice image; Acquiring a plurality of second slice images, wherein the plurality of second slice images represent tissue slices of a plurality of subjects, the plurality of subjects having the same properties as the subjects constituting the administration group and being in the same breeding environment, and constituting a control group of the candidate substance to which no drug is administered; extracting second feature quantities from each of the plurality of second slice images; acquiring third feature quantity information related to a third feature quantity extracted from a plurality of third slice images showing tissue slices of a plurality of subjects constituting a control group to which the candidate substance was not administered in a past evaluation test; and Based on the first feature amount, the second feature amount, and the third feature amount information, it is determined whether or not there is an abnormality in the tissue slice shown in the first slice image.