Cell classification device, cell classification method, and program

Abstract

In order to classify sample cells as being benign or malignant without being limited to a specific procedure, a cell classification device (1) comprises: an acquisition means (10) that acquires an image including sample cells as a subject; a determination means (20) that determines whether each sample cell is an individual cell or a cluster of cells; and a classification means (30) that performs at least one of classifying a sample cell determined to be an individual cell as being benign or malignant and classifying a sample cell determined to be a cluster of cells as being benign or malignant.

Description

Technical Field

[0001] The present invention relates to a cell sorting device, a cell sorting method, and a program.

Background Art

[0002] When a tumor is suspected by an examination such as an X-ray or an X-ray CT, a definitive diagnosis is generally made by a pathological diagnosis. For example, a part of a visceral tissue of the human body is collected as a specimen by puncture or the like, and the diagnosis is made by microscopic observation of the specimen by an expert. In the process of specimen collection, it may be necessary to quickly observe whether the specimen contains tissue cells suitable for pathological diagnosis in the treatment room where the specimen collection is being performed. This is because if the specimen does not contain tissue cells suitable for pathological diagnosis, it is necessary to collect the specimen again. However, specimen collection by puncture or the like is invasive, and it is not preferable to perform specimen collection many times.

[0003] Intraoperative rapid cytodiagnosis (ROSE, Rapid On-Site Evaluation) is a technique for quickly observing in the treatment room where the specimen is collected whether the collected specimen contains cells suitable for pathological diagnosis and whether the specimen contains malignant cells. For example, it is used when collecting a specimen by puncture using an endoscope for a respiratory organ.

[0004] However, specialized knowledge and experience are required to determine whether the collected cells are benign or malignant, and the human resources of such skilled personnel are limited. For this reason, although ROSE has been reported to be useful in reducing the number of punctures and the risk of complications, there is a problem that its implementation is limited. Therefore, attempts have been made to enable a non-expert to make a diagnosis close to that of an expert by assisting such a determination with a mechanical device.

[0005] For example, Patent Document 1 discloses a technique for calculating morphological features of cells from an image of a specimen stained with a molecular target dye that visualizes a predetermined target molecule and determining the degree of cell abnormality. This includes determining the normality / abnormality of each of the recognized cell or cell cluster regions. Further, Patent Document 2 discloses determining whether it is an isolated single cell (isolated scattered cell) or a cell cluster composed of a plurality of cells, and determining the normality / abnormality of an isolated scattered cell based on at least one of a shape feature parameter related to the image of the cell represented in the whole specimen VS the image of the cell and a video feature parameter (such as a luminance value) related to the image of the cell, and performing the abnormality determination process of the cell cluster by changing the table used for the determination.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0007] However, the technique described in Patent Document 1 performs counterstaining to stain cell nuclei on a specimen and evaluates whether a desired antigen is expressed around the cell nuclei stained by this counterstaining, and is a technique limited to observing whether the antigen is expressed or not. Further, the technique described in Patent Document 2 is a technique limited to an analysis method of a virtual slide image of the specimen configured by combining a plurality of microscope images obtained by photographing each time the objective lens and the specimen are relatively moved in a direction orthogonal to the optical axis.

[0008] One aspect of the present invention has been made in view of the above problems, and an example of its object is to provide a technique that enables improvement in the accuracy of classifying the malignancy of sample cells without being limited to specific purposes or methods.

Means for Solving the Problems

[0009] A cell classification apparatus according to one aspect of the present invention includes an acquisition unit that acquires an image including a sample cell as a subject, a determination unit that determines whether each of the sample cells is an individual cell or a mass of cells, and a classification unit that performs at least one of a process of classifying the malignancy of the sample cells determined to be individual cells and a process of classifying the malignancy of the sample cells determined to be masses of cells.

[0010] A cell classification method according to one aspect of the present invention includes at least one processor acquiring an image including a sample cell as a subject, determining whether each of the sample cells is an individual cell or a mass of cells, and performing at least one of a process of classifying the malignancy of the sample cells determined to be individual cells and a process of classifying the malignancy of the sample cells determined to be masses of cells.

[0011] A cell classification program according to one aspect of the present invention causes a computer to execute a process of acquiring an image including a sample cell as a subject, a process of determining whether each of the sample cells is an individual cell or a mass of cells, and a process of performing at least one of a process of classifying the malignancy of the sample cells determined to be individual cells and a process of classifying the malignancy of the sample cells determined to be masses of cells.

Effects of the Invention

[0012] According to one aspect of the present invention, the malignancy of sample cells can be classified without being limited to a specific method.

Brief Description of the Drawings

[0013]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

BEST MODE FOR CARRYING OUT THE INVENTION

[0014] 〔Exemplary Embodiment 1〕 The first exemplary embodiment of the present invention will be described in detail with reference to the drawings. This exemplary embodiment is a basic form for the exemplary embodiments described later.

[0015] (Configuration of the cell sorter 1) The configuration of the cell classification device 1 according to this exemplary embodiment will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the configuration of the cell classification device 1 according to Exemplary Embodiment 1. The cell classification device 1 is a device that classifies whether cells (specimen cells) contained in a specimen are benign cells or malignant cells (classifies benign and malignant). Pathological diagnosis refers to observing a specimen collected from the human body under a microscope and diagnosing the presence or absence of a lesion and the type of the lesion. The cell classification device 1 is, as an example, a device for performing intraoperative rapid cytodiagnosis during specimen collection by puncture using an endoscope of a respiratory organ.

[0016] As shown in FIG. 1, the cell classification device 1 includes an acquisition unit 10, a determination unit 20, and a classification unit 30. The acquisition unit 10 acquires an image including specimen cells as a subject. In this exemplary embodiment, the specimen cells are cells contained in a specimen, and the specimen usually contains a plurality of specimen cells. The specimen may be, as an example, a specimen of an organ tissue collected by puncture. Puncture is a specimen collection method in which a doctor inserts a sampling needle into an organ tissue to collect cells of the organ tissue. The puncture may be performed by the doctor while using a respiratory endoscope. Thereby, for example, a tissue suspected of having a lung tumor can be collected. The collected organ tissue is, for example, made into a specimen by a medical technician, and an image (microscope camera image) including specimen cells (hereinafter, also simply referred to as "cells") is taken using a microscope with a digital camera. The medical technician records the taken image in a memory (not shown) of the cell classification device 1, for example. The acquisition unit 10 acquires the image recorded in the memory as digital data. The acquisition unit 10 is a form of the "acquisition means" described in the claims.

[0017] The medical technician prepares one or more slide specimens using the collected specimen, stains the cells, air-dries them, and then takes one or more images using a microscope. These operations are generally performed operations, and it is possible to determine at this point whether sufficient cells have been collected. If sufficient cells have not been collected, puncture can be performed again to collect a specimen.

[0018] The determination unit 20 determines whether each of the sample cells is an individual cell or a clump of cells. An individual cell is a cell that exists as one independent cell. A clump of cells (hereinafter also referred to as a "cell clump") is a cell in which a plurality of cells are gathered in a densely packed state. Also, cells in which a plurality of individual cells are photographed overlapping in the depth direction are also included in the cell clumps. The determination unit 20 is a form of the "determination means" described in the claims.

[0019] In order to determine whether it is an individual cell or a clump of cells, first, the cells are detected. The method of detecting cells will be described with reference to the drawings. FIG. 2 is a schematic diagram showing an example of a method of detecting cells from an image 201 including sample cells as a subject. First, the determination unit 20 generates a staining intensity image 202 from the image 201 of FIG. 2 by regression analysis. Specifically, the determination unit 20 sets the brightness of the unstained background region (including regions with weak staining such as red blood cells) to 0, and generates a black-and-white image 202 in which the stained region (staining region) is assigned a gradation of brightness from 1 to 255 for each pixel according to the staining intensity (Reference: US2022 / 0028068 A1, applicant NEC Laboratories America). This image 202 is the staining intensity image.

[0020] The determination unit 20 may determine whether it is an individual cell or a clump of cells with reference to at least one of the size and circularity of the sample cells. As an example, the determination unit 20 determines whether the area of a continuous staining region in the staining intensity image is equal to or greater than a predetermined threshold. The determination unit 20 determines (detects) a region where the area of the continuous staining region is equal to or greater than the predetermined threshold as a cell clump. The image 203 in FIG. 2 is an image with only cell clumps remaining.

[0021] On the other hand, the determination unit 20 determines that a region where the area of the continuous staining region is smaller than the predetermined threshold is a candidate region for individual cells. The determination unit 20 may finally detect individual cells from among these candidate regions using this region as a candidate. The image 204 in FIG. 2 is an image with only candidate regions for individual cells remaining.

[0022] The reason for determining whether the cells are individual cells or clump cells is that in the case of clump cells, the criteria for determination are different from those for individual cells. Therefore, when performing ROSE using the cell classification device 1, it is preferable to detect the cells, determine whether the detected cells are individual cells or clump cells, and determine whether the cells are benign or malignant according to whether they are individual cells or clump cells.

[0023] The classification unit 30 performs at least one of the processes of classifying the detected cells determined to be individual cells as benign or malignant and classifying the detected cells determined to be clump cells as benign or malignant.

[0024] For example, the classification unit 30 may perform benign / malignant classification only on the cells determined to be individual cells. The machine learning model for performing benign / malignant classification is trained based on cells that can be visually determined as benign or malignant on an individual cell basis, in other words, cells that can be labeled as benign or malignant. Therefore, the accuracy of benign / malignant classification of individual cells is better than that of clump cells. Thus, if benign / malignant classification is applied only to the cells determined to be individual cells, the classification accuracy is improved compared to the case of performing benign / malignant classification on individual cells and clump cells together. In addition, since there is no need to perform benign / malignant classification of clump cells, the time required for the benign / malignant classification process can be shortened compared to the case of performing benign / malignant classification including clump cells.

[0025] In addition, the classification unit 30 may perform benign / malignant classification only on the cells determined to be clump cells. Alternatively, the classification unit 30 may perform benign / malignant classification on both the detected cells determined to be individual cells and the detected cells determined to be clump cells. Note that the "benign / malignant classification" is not limited to the classification of benign or malignant, and may be, for example, calculating the probability of malignancy.

[0026] The classification unit 30 classifies either, or both, one or more individual cells and one or more mass cells into benign and malignant. The classification unit 30 may further aggregate the results of the benign and malignant classification. For example, the classification unit 30 may aggregate data such as the number of cells determined to be benign and the number of cells determined to be malignant, respectively. The classification unit 30 may refer to the aggregated results to determine whether the specimen is benign or malignant. The determination of the specimen can be made, for example, by comprehensively aggregating the results of a plurality of cells subjected to benign and malignant classification. The classification unit 30 is a form of the "classification means" described in the claims.

[0027] Note that in FIG. 1, the acquisition unit 10, the determination unit 20, and the classification unit 30 are described as being physically arranged together in one housing, but this is not necessarily required. That is, these units may be physically distributed and arranged in a plurality of separated housings, and these units may be connected to each other by wire or wirelessly so as to be able to communicate information. Also, at least a part of these units may be arranged on the cloud. This is the same in the following exemplary embodiments.

[0028] As described above, in the cell classification apparatus 1 according to this exemplary embodiment, an acquisition unit 10 that acquires an image including a specimen cell as a subject, a determination unit 20 that determines whether each of the specimen cells is an individual cell or a mass cell, and a classification unit 30 that performs at least one of a process of classifying the specimen cells determined to be individual cells into benign and malignant and a process of classifying the specimen cells determined to be mass cells into benign and malignant are provided. Therefore, according to the cell classification apparatus 1 according to this exemplary embodiment, the effect of being able to improve the accuracy of classifying the specimen cells into benign and malignant without being limited to a specific purpose or method can be obtained. Also, when performing benign and malignant classification only on the cells determined to be individual cells, the effect of improving the classification accuracy compared to the case of classifying individual cells and mass cells together into benign and malignant can be obtained.

[0029] ( Cell classification method S1 of the flow) Regarding the flow of the cell classification method S1 according to the present exemplary embodiment, it will be described with reference to FIG. 3. FIG. 3 is a flowchart showing the flow of the cell classification method S1. As shown in FIG. 3, the cell classification method S1 includes steps S11 to S13.

[0030] Step S11 is a step in which at least one processor (for example, the acquisition unit 10) acquires an image including the specimen cells as a subject. The specimen cells and the image are as described in the configuration of the cell classification apparatus 1.

[0031] Step S12 is a step in which at least one processor (for example, the determination unit 20) determines whether each of the specimen cells is an individual cell or a clump of cells. The individual cells and the clump of cells, and the method of determination are as described in the configuration of the cell classification apparatus 1.

[0032] Step S13 is a step in which at least one processor (for example, the classification unit 30) performs at least one of classifying the specimen cells determined to be individual cells into benign and malignant, and classifying the specimen cells determined to be clumps of cells into benign and malignant. The classification of benign and malignant is as described in the configuration of the cell classification apparatus 1.

[0033] As described above, in the cell classification method S1 according to the present exemplary embodiment, a configuration including acquiring an image including the specimen cells as a subject, determining whether each of the specimen cells is an individual cell or a clump of cells, classifying the specimen cells determined to be individual cells into benign and malignant, and classifying the specimen cells determined to be clumps of cells into benign and malignant, and performing at least one of these is adopted. Therefore, according to the cell classification method S1 according to the present exemplary embodiment, the effect of being able to improve the accuracy of classifying the specimen cells into benign and malignant without being limited to a specific purpose or method can be obtained.

[0034] 〔Exemplary Embodiment 2〕 The second exemplary embodiment of the present invention will be described in detail with reference to the drawings. Note that the same reference numerals are used for the components having the same functions as those described in the first exemplary embodiment, and the description thereof will be omitted as appropriate.

[0035] (Configuration of cell sorting device 1A) 4 is a block diagram showing the configuration of a cell sorting device 1A according to exemplary embodiment 2. As shown in the figure, the cell sorting device 1A includes an acquisition section 10, a determination section 20A, a classification section 30, an output section 40, and a control section 50.

[0036] The functions of the acquisition unit 10 are similar to those of the acquisition unit 10 described in exemplary embodiment 1, but in this exemplary embodiment, the acquisition unit 10 acquires images of organ tissue that are pre-recorded in the memory 52 of the control unit 50.

[0037] The judgment unit 20A includes a trained judgment model 21. The trained judgment model 21 may be a judgment model trained using a teacher image showing an individual cell and a teacher image showing a clustered cell. FIG. 5 is an example of teacher data used to construct the trained judgment model 21. As shown in the figure, the trained judgment model 21 is trained using a plurality of images (teacher data) of individual cells and a plurality of images (teacher data) of clustered cells, and can output whether a cell in an image is an individual cell or a clustered cell. The judgment unit 20A inputs an image to the trained judgment model 21, and judges whether each cell in the image is an individual cell or a clustered cell by referring to the output of the trained judgment model 21. As an example, the trained judgment model 21 can be a CNN (Convolution Neural Network). In addition, a non-neural network type model such as a random forest or a support vector machine may be used.

[0038] Returning to FIG. 4, after detecting each of the specimen cells as individual cell candidates or mass cell candidates, the determination unit 20A may determine the specimen cells detected using the learned determination model 21. As an example, the determination unit 20A may detect candidate regions for individual cells and mass cell candidates using the method described in the first exemplary embodiment. Then, the determination unit 20A may perform the following processing, for example, on the candidate regions for individual cells to detect individual cell candidates. That is, the determination unit 20A searches for a local maximum value of brightness within a predetermined range, for example, a circular region with a diameter of n (mm) or more and less than m (mm). The point (pixel) of the local maximum value of brightness is taken as the cell center. Subsequently, the determination unit 20A extracts, by Hough transform, those regions having a circularity of a predetermined brightness or more from the cell center. The determination unit 20A detects the extracted regions as individual cell candidates. As described above, usually, a plurality of individual cell candidates and mass cell candidates are detected from one specimen. The determination unit 20A finally determines whether each of the plurality of individual cell candidates and mass cell candidates detected as described above is an individual cell or a mass cell using the learned determination model 21.

[0039] Similar to the classification unit 30 described in the first exemplary embodiment, the classification unit 30 performs at least one of the processing of performing a benign / malignant classification on the cells determined to be individual cells and the processing of performing a benign / malignant classification on the cells determined to be mass cells. Further, the classification unit 30 may transmit the result of the benign / malignant classification to the output unit.

[0040] The output unit 40 may output the result of the benign / malignant classification by the classification unit 30 to the outside. The output benign / malignant classification result may be displayed on, for example, a display device (not shown).

[0041] The control unit 50 controls the entire cell classification device 1A. The control unit 50 includes at least one processor and a memory 52. ​​The processor 51 can be configured using a general-purpose processor such as at least one MPU (Micro Processing Unit) or CPU (Central Processing Unit). The memory 52 may include multiple types of memory such as ROM (Read Only Memory) and RAM (Random Access Memory). As an example, the processor 51 realizes the functions of the acquisition unit 10, the judgment unit 20A, the classification unit 30, and the output unit 40 by expanding various control programs recorded in the ROM of the memory 52 into the RAM and executing them. The processor 51 may also include a processor configured with an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), a PLD (Programmable Logic Device), or the like.

[0042] (Effects of cell sorting device 1A) As described above, the cell sorting device 1A according to this exemplary embodiment employs a configuration in which the judgment unit 20A judges whether each cell is an individual cell or an aggregated cell using the learned judgment model 21 that has been trained using a teacher image showing an individual cell and a teacher image showing an aggregated cell. Therefore, in addition to the effects of the cell sorting device 1 according to the exemplary embodiment 1, the cell sorting device 1A according to this exemplary embodiment can efficiently judge individual cells, thereby efficiently classifying individual cells into benign or malignant cells.

[0043] In the cell classification method, the determination of whether each of the specimen cells is an individual cell or an aggregated cell may be performed using a determination model 21 trained using a teacher image showing an individual cell and a teacher image showing an aggregated cell. This makes it possible to obtain the same effect as that of the cell classification device 1A described above.

[0044] Exemplary embodiment 3 The third exemplary embodiment of the present invention will be described in detail with reference to the drawings. Note that the same reference numerals are given to the parts having the same functions as those described in the first exemplary embodiment, and the description thereof will not be repeated.

[0045] (Configuration of cell sorting device 1B) 6 is a block diagram showing the configuration of a cell classification device 1B according to exemplary embodiment 3. As shown in the figure, the cell classification device 1B includes an acquisition unit 10, a determination unit 20, a classification unit 30A, an output unit 40, and a control unit 50. The acquisition unit 10, the output unit 40, and the control unit 50 have the same functions as the acquisition unit 10, the output unit 40, and the control unit 50 described in exemplary embodiment 2, and therefore a description thereof will be omitted. Note that the cell classification device 1B may include the determination unit 20A described in exemplary embodiment 2 instead of the determination unit 20.

[0046] In the present exemplary embodiment 3, the classification unit 30A includes a trained classification model 31. The trained classification model 31 may be a classification model trained using at least one of a teacher image showing individual benign cells and malignant cells and a teacher image showing clumped benign cells and malignant cells. As an example, the trained classification model 31 may use a CNN (Convolution Neural Network). A non-neural network model such as a random forest or a support vector machine may also be used. The classification unit 30A, like the classification unit 30 described in the exemplary embodiment 1, performs at least one of a process of performing benign / malignant classification on cells determined to be individual cells and a process of performing benign / malignant classification on cells determined to be clumped cells, using the trained classification model 31. That is, the classification unit 30A inputs an image of a cell to be classified into the trained classification model 31, and performs benign / malignant classification of the cell by referring to the output.

[0047] The learned classification model 31 may be a learned individual cell classification model learned using teacher images indicating benign and malignant individual cells. In that case, the classification unit 30A acquires only the cells determined by the determination unit 20 to be individual cells from the determination unit 20 and performs benign / malignant classification using the learned individual cell classification model. FIG. 7 is a schematic diagram showing an example of constructing a learned individual cell classification model using teacher images indicating benign and malignant individual cells. As shown in FIG. 7, the user inputs an image indicating a benign individual cell into the individual cell classification model to learn the characteristics of the benign individual cell, and inputs an image indicating a malignant individual cell to learn the characteristics of the malignant individual cell. In this way, the learned individual cell classification model 31 can be constructed.

[0048] Further, the learned classification model 31 may be configured to include two learned classification models, namely, the learned individual cell classification model 31A learned as described above and a learned mass cell classification model 31B learned using teacher images indicating mass-like benign and malignant cells as shown in FIG. 8 (not shown). As shown in FIG. 8, the user inputs an image indicating a benign mass cell into the mass cell classification model to learn the characteristics of the benign mass cell, and inputs an image indicating a malignant mass cell to learn the characteristics of the malignant mass cell. In this way, the learned mass cell classification model 31B can be constructed.

[0049] Even when the learned classification model 31 is configured to include two models, i.e., the learned individual cell classification model 31A and the learned mass cell classification model 31B, the classification unit 30A may perform benign / malignant classification using the learned individual cell classification model 31A only for the cells determined to be individual cells. Further, the classification unit 30A may perform benign / malignant classification using the learned mass cell classification model 31B only for the cells determined to be mass cells.

[0050] The classification unit 30A may perform benign / malignant classification using both the learned individual cell classification model 31A and the learned mass cell classification model 31B. In that case, the classification unit 30A refers to both the output result of the learned individual cell classification model 31A and the output result of the learned mass cell classification model 31B to perform benign / malignant classification of the specimen cells. For example, the classification unit 30A may perform benign / malignant classification by referring to the data of individual cells classified as malignant and the data of mass cells classified as malignant. For example, the classification unit 30A may calculate the classification result using the following formula (1).

[0051] [Number]

[0052] In the above formula (1), p is the probability that the specimen is malignant, w1 is the weight of the output of the learned individual cell classification model 31A, w2 is the weight of the output of the learned mass cell classification model 31B, Σp1 is the sum of the probabilities of being malignant for individual cells, Σp2 is the sum of the probabilities of being malignant for mass cells, n is the number of individual cells, m is the number of mass cells.

[0053] Even for mass cells, there may be cases where information for determining whether the specimen is benign or malignant is included. Therefore, by performing benign / malignant classification of mass cells, such information can also be obtained. However, generally, it is known that the determination of the benign / malignancy of mass cells is more difficult than that of individual cells, that is, the classification accuracy is lower. Therefore, w2 may be set smaller than w1.

[0054] Also, the classification unit 30A may transmit at least any one of the data of individual cells classified as malignant, the data of mass cells classified as malignant, and the result of benign / malignant classification of the specimen to the output unit 40.

[0055] FIG. 9 is a schematic diagram showing an example of a procedure for generating input data for performing benign / malignant classification using both the learned individual cell classification model 31A and the learned mass cell classification model 31B.

[0056] First, as shown in FIG. 9, the acquisition unit 10 acquires an image 901 including a specimen cell as a subject. Next, the determination unit 20 generates a staining intensity image 902 from the image 901. Next, the determination unit 20 generates a mass cell candidate image 903 and a candidate region image 905 of individual cells from the staining intensity image 902. Further, the determination unit 20 detects mass cells from the mass cell candidate image 903 and detects individual cell candidates from the candidate region image 905 of individual cells. An example of these methods is as described in the exemplary embodiment 1. Next, the determination unit 20 may generate an image 904 obtained by cutting out the individual cell from the image of the cell determined to be an individual cell to a predetermined size. In addition, the determination unit 20 may cut out an image 907 of an arbitrary size (an arbitrary range from the center of the mass or a bounding rectangle, etc.) including the cell from the image 906 of the cell determined (detected) to be a mass cell and generate an image normalized to a predetermined size. This is because the images input to the learned classification model need to be unified to a predetermined size. The determination unit 20 transmits the images generated in this way to the classification unit 30A. The classification unit 30A inputs the acquired images to the learned individual cell classification model 31A or the learned mass cell classification model 31B according to their types, and performs benign / malignant classification with reference to these outputs.

[0057] (Effect of the cell classification device 1B) As described above, in the cell sorting device 1B according to the present exemplary embodiment, the sorting unit 30A is configured to include a sorting model trained using at least one of a teacher image showing benign and malignant cells of individual cells and a teacher image showing clumped benign and malignant cells. Therefore, according to the cell sorting device 1B according to the present exemplary embodiment, particularly when performing benign / malignant classification using both the trained individual cell sorting model 31A and the trained clumped cell sorting model 31B, in addition to the effects of the cell sorting device 1 according to the exemplary embodiment 1, it is possible to perform benign / malignant classification including information contained in clumped cells, and it is possible to achieve an effect of improving the accuracy of benign / malignant classification of samples.

[0058] In the cell classification method, the benign / malignant classification may be performed using a classification model trained using at least one of a teacher image showing individual benign cells and malignant cells and a teacher image showing clumped benign cells and malignant cells. In particular, when performing benign / malignant classification using both the trained individual cell classification model 31A and the trained clumped cell classification model 31B, the same effect as that of the above-mentioned cell classification device 1B can be obtained.

[0059] [Software implementation example] A part or all of the functions of the cell sorting devices 1, 1A, 1B may be realized by hardware such as an integrated circuit (IC chip), or may be realized by software.

[0060] In the latter case, the cell sorting devices 1, 1A, 1B are realized, for example, by a computer that executes instructions of a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as computer C) is shown in FIG. 10. The computer C has at least one processor C1 and at least one memory C2. The memory C2 stores a program P for operating the computer C as the cell sorting devices 1, 1A, 1B. In the computer C, the processor C1 reads and executes the program P from the memory C2, thereby realizing each function of the cell sorting devices 1, 1A, 1B.

[0061] As the processor C1, for example, a CPU (Central Processing Unit), GPU (Graphic Processing Unit), DSP (Digital Signal Processor), MPU (Micro Processing Unit), FPU (Floating point number Processing Unit), PPU (Physics Processing Unit), TPU (Tensor Processing Unit), quantum processor, microcontroller, or a combination thereof can be used. As the memory C2, for example, a flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or a combination thereof can be used.

[0062] Note that the computer C may further include a RAM (Random Access Memory) for expanding the program P during execution and temporarily storing various data. The computer C may further include a communication interface for transmitting and receiving data to and from other devices. The computer C may further include an input / output interface for connecting input / output devices such as a keyboard, mouse, display, and printer.

[0063] Also, the program P can be recorded on a non-transitory tangible recording medium M readable by the computer C. As such a recording medium M, for example, a tape, disk, card, semiconductor memory, or programmable logic circuit can be used. The computer C can obtain the program P via such a recording medium M. Also, the program P can be transmitted via a transmission medium. As such a transmission medium, for example, a communication network or broadcast wave can be used. The computer C can also obtain the program P via such a transmission medium.

[0064] 〔Supplementary Note 1〕 The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope indicated in the claims. For example, embodiments obtained by appropriately combining the technical means disclosed in the above-described embodiments are also included in the technical scope of the present invention.

[0065] 〔Supplementary Note 2〕 Some or all of the above-described embodiments may also be described as follows. However, the present invention is not limited to the aspects described below.

[0066] (Supplementary Note 1) An acquisition means for acquiring an image including a specimen cell as a subject, a determination means for determining whether each of the specimen cells is an individual cell or a clump of cells, a process for performing a benign / malignant classification on the specimen cells determined to be individual cells, and a classification means for performing at least one of the processes for performing a benign / malignant classification on the specimen cells determined to be clumps of cells. A cell classification device characterized by that. According to the above configuration, it is possible to perform a benign / malignant classification of specimen cells without being limited to a specific method.

[0067] (Supplementary Note 2) The classification means performs a benign / malignant classification only on the specimen cells determined to be individual cells. The cell classification device according to Supplementary Note 1. According to the above configuration, by performing a benign / malignant classification excluding clump cells, the accuracy of the benign / malignant classification of the specimen can be made higher.

[0068] (Supplementary Note 3) The classification means performs a benign / malignant classification on both the specimen cells determined to be individual cells and the specimen cells determined to be clumps of cells. The cell classification device according to Supplementary Note 1. According to the above configuration, even if it appears to be a clump cell, by performing a benign / malignant classification including individual cells that can be subjected to a benign / malignant classification, the accuracy of the benign / malignant classification of the specimen can be made higher.

[0069] (Supplementary Note 4) The classification means is the cell classification device described in Supplementary Note 3 that performs benign / malignant classification with reference to the data of the individual cells classified as malignant and the data of the mass cells classified as malignant. According to the above configuration, it becomes possible to perform benign / malignant classification including the information contained in the mass cells, and the accuracy of the benign / malignant classification of the specimen can be made higher.

[0070] (Supplementary Note 5) The determination means is the cell classification device described in any one of Supplementary Notes 1 to 4 that determines whether the cell is an individual cell or a mass cell with reference to at least either the size or the circularity of the specimen cell. According to the above configuration, it is possible to effectively determine whether the cell is an individual cell or a mass cell.

[0071] (Supplementary Note 6) The determination means is the cell classification device described in any one of Supplementary Notes 1 to 4 that includes a determination model learned using a teacher image indicating an individual cell and a teacher image indicating a mass cell. According to the above configuration, it is possible to determine individual cells and mass cells using the learned determination model, and the accuracy of the benign / malignant classification of the specimen can be made higher.

[0072] (Supplementary Note 7) The determination means is the cell classification device described in Supplementary Note 6 that, after detecting each of the specimen cells as an individual cell candidate or a mass cell candidate, determines the detected specimen cells using the learned determination model. According to the above configuration, it becomes possible to perform benign / malignant classification including the information contained in the mass cells, and the accuracy of the benign / malignant classification of the specimen can be made higher.

[0073] (Supplementary Note 8) The classification means is the cell classification device described in any one of Supplementary Notes 1 to 7 that includes a classification model learned using at least either a teacher image indicating a benign cell and a malignant cell of an individual cell or a teacher image indicating a benign cell and a malignant cell of a mass cell. According to the above configuration, benign and malignant cells can be classified using a learned classification model, thereby improving the accuracy of classifying samples as benign or malignant.

[0074] (Appendix 9) The cell classification device of Appendix 8, wherein the determination means generates an image of a predetermined size that includes an individual cell from an image of the specimen cell that has been determined to be an individual cell, or an image of an arbitrary size that includes the clumped cell from an image of the specimen cell that has been determined to be a clumped cell, and standardizes the image to the predetermined size. According to the above configuration, it is possible to generate images that can be input into a trained classification model.

[0075] (Appendix 10) 10. The cell classification device according to any one of claims 1 to 9, further comprising an output means for outputting the benign / malignant classification result obtained by the classification means. According to the above configuration, the classification result can be output to a display device.

[0076] (Appendix 11) A cell classification method including at least one processor performing at least one of the following: acquiring an image including specimen cells as a subject; determining whether each of the specimen cells is an individual cell or a clump of cells; classifying the specimen cells determined to be individual cells as benign or malignant; and classifying the specimen cells determined to be clumps of cells as benign or malignant. According to the above configuration, the same effect as that of Supplementary Note 1 can be obtained.

[0077] (Appendix 12) The cell classification method according to claim 11, wherein the classification into benign or malignant is performed only on the specimen cells that have been determined to be individual cells. According to the above configuration, the same effect as that of Supplementary Note 2 can be obtained.

[0078] (Appendix 13) The method for classifying into benign and malignant cells is the method for classifying cells according to Supplementary Note 11, which comprises classifying into benign and malignant cells for both the specimen cells determined to be individual cells and the specimen cells determined to be clump-shaped cells respectively. According to the above configuration, an effect similar to the effect of Supplementary Note 3 can be obtained.

[0079] (Supplementary Note 14) The method for classifying into benign and malignant cells is the method for classifying cells according to Supplementary Note 13, which comprises classifying into benign and malignant cells with reference to the data of the individual cells classified as malignant and the data of the clump-shaped cells classified as malignant. According to the above configuration, an effect similar to the effect of Supplementary Note 4 can be obtained.

[0080] (Supplementary Note 15) The determination is the method for classifying cells according to any one of Supplementary Notes 11 to 14, which comprises determining whether the specimen cells are individual cells or clump-shaped cells with reference to at least either the size or the circularity of the specimen cells. According to the above configuration, an effect similar to the effect of Supplementary Note 5 can be obtained.

[0081] (Supplementary Note 16) The determination is the method for classifying cells according to any one of Supplementary Notes 11 to 14, which comprises determining using a determination model learned using a teacher image indicating individual cells and a teacher image indicating clump-shaped cells. According to the above configuration, an effect similar to the effect of Supplementary Note 6 can be obtained.

[0082] (Supplementary Note 17) The determination is the method for classifying cells according to Supplementary Note 16, which comprises detecting each of the specimen cells as an individual cell candidate or a clump-shaped cell candidate, and then determining the specimen cells detected using the learned determination model. According to the above configuration, an effect similar to the effect of Supplementary Note 7 can be obtained.

[0083] (Supplementary Note 18) The step of performing the benign / malignant classification is to perform the benign / malignant classification using a classification model learned using at least one of teacher images indicating individual benign and malignant cells and teacher images indicating mass-like benign and malignant cells, according to the cell classification method described in any one of Appendices 11 to 17. According to the above configuration, an effect similar to the effect of Appendix 8 can be obtained.

[0084] (Appendix 19) The step of making the determination is to generate an image by cutting out an image of a predetermined size including the individual cell from an image of the specimen cell determined to be an individual cell, or an image of an arbitrary size including the mass-like cell from an image of the specimen cell determined to be a mass-like cell and normalizing the image to the predetermined size, and making the determination using the image, according to the cell classification method described in Appendix 18. According to the above configuration, an effect similar to the effect of Appendix 9 can be obtained.

[0085] (Appendix 20) The cell classification method according to any one of Appendices 11 to 19, further including outputting the result of the benign / malignant classification. According to the above configuration, an effect similar to the effect of Appendix 10 can be obtained.

[0086] (Appendix 21) A cell classification program for causing a computer to execute a process of acquiring an image including a specimen cell as a subject, a process of determining whether each of the specimen cells is an individual cell or a mass-like cell, a process of performing a benign / malignant classification on the specimen cell determined to be an individual cell, and a process of performing a benign / malignant classification on the specimen cell determined to be a mass-like cell, and at least one of the processes.

[0087] [Appendix Item 3] Some or all of the above-described embodiments can also be expressed as follows. A cell classification device comprising at least one processor, the processor performing an acquisition process of acquiring an image including a specimen cell as a subject, a determination process of determining whether each of the specimen cells is an individual cell or a clump of cells, and performing at least any one of a process of classifying the specimen cells determined to be individual cells into benign and malignant, and a classification process of classifying the specimen cells determined to be clumps of cells into benign and malignant.

[0088] Note that this cell classification device may further include a memory, and a program for causing the processor to execute the acquisition process, the determination process, and the classification process may be stored in this memory. Further, this program may be recorded on a non-transitory tangible recording medium readable by a computer.

Explanation of Signs

[0089] 1, 1A, 1B… Cell classification device 10… Acquisition unit 20, 20A… Determination unit 21… Learned determination model 30, 30A… Classification unit 31… Learned classification model 40… Output unit 50… Control unit 51… Processor 52… Memory

Claims

1. An acquisition means for acquiring an image including a subject somatic cell; A determination means for determining whether each of the subject somatic cells is an individual cell or a mass of cells; A classification means for performing at least one of a process of classifying the subject somatic cells determined to be individual cells as benign or malignant and a process of classifying the subject somatic cells determined to be masses of cells as benign or malignant A cell classification device comprising the above.

2. The classification means performs benign / malignant classification only on the subject somatic cells determined to be individual cells The cell classification device according to Claim 1.

3. The classification means performs benign / malignant classification on both the subject somatic cells determined to be individual cells and the subject somatic cells determined to be masses of cells, respectively, with reference to the data of the individual cells classified as malignant and the data of the masses of cells classified as malignant The cell classification device according to Claim 1.

4. The determination means determines whether the subject somatic cell is an individual cell or a mass of cells with reference to at least one of the size and circularity of the subject somatic cell The cell classification device according to any one of Claims 1 to 3.

5. The determination means includes a determination model learned using a teacher image indicating an individual cell and a teacher image indicating a mass of cells The cell classification device according to any one of Claims 1 to 3.

6. After the determination means detects each of the subject somatic cells as an individual cell candidate or a mass of cell candidate, the determination means determines the detected subject somatic cell using the learned determination model The cell classification device according to Claim 5.

7. The classification means includes a classification model learned using at least one of a teacher image indicating a benign cell and a malignant cell of an individual cell and a teacher image indicating a benign cell and a malignant cell of a mass of cells The cell classification device according to any one of Claims 1 to 3.

8. The determination means generates an image obtained by cutting out an image of a predetermined size including the individual cell from an image of the subject somatic cell determined to be an individual cell, or an image of an arbitrary size including the mass of cells from an image of the subject somatic cell determined to be a mass of cells and normalizing the image to the predetermined size The cell classification device according to Claim 7.

9. At least one processor Acquires an image including a subject somatic cell; Determines whether each of the subject somatic cells is an individual cell or a mass of cells; Performing at least either classifying the sample cells determined to be individual cells as benign or malignant, and classifying the sample cells determined to be clustered cells as benign or malignant A cell classification method including the above

10. Causing a computer to perform a process of acquiring an image including the sample cells as a subject, perform a process of determining whether each of the sample cells is an individual cell or a clustered cell, perform at least either a process of classifying the sample cells determined to be individual cells as benign or malignant, and a process of classifying the sample cells determined to be clustered cells as benign or malignant A cell classification program for causing the above to be executed

Images