WO2021085648A1 - Cell evaluation device, program, and cell evaluation method - Google Patents

Abstract

A cell evaluation device comprising: an evaluation condition acquisition unit for acquiring a cell evaluation condition selected by a user; an evaluation indicator selection unit for selecting an evaluation indicator which is obtained from an image including the cell on the basis of the evaluation condition acquired by the evaluation condition acquisition unit, and in which a value showing the accuracy of separating a positive example and a negative example of the evaluation condition is equal to or larger than a predetermined value, and specifying a time suitable for the acquisition of an image to be used for the calculation of the evaluation indicator; and an evaluation unit for evaluating the cell with respect to the evaluation condition on the basis of the evaluation indicator selected by the evaluation indicator selection unit in the image acquired at the specified time.

Description

Cell evaluation device, program, and cell evaluation method

The present invention relates to a cell evaluation device, a program, and a cell evaluation method.
The present application claims priority based on Japanese Patent Application No. 2019-198412 filed in Japan on October 31, 2019, the contents of which are incorporated herein by reference.

In recent years, a method for inducing differentiation from stem cells to differentiated cells has been established, and a technique for stable cell culture is known. Patent Document 1 describes a technique for culturing cells by a method for inducing terminal differentiation of human functional cells based on tissue reconstruction.

International Publication No. 2013/047639

Here, cells may be evaluated under various evaluation conditions. The evaluation conditions include, for example, quality, yield, speed of culture, and the like. In this case, it is preferable that the cells are evaluated based on the evaluation index according to the evaluation conditions.

Therefore, an object of the present invention is to provide a cell evaluation device, a program, and a cell evaluation method for evaluating the state of cells under various evaluation conditions.

It is a figure which shows an example of the cell evaluation system which concerns on 1st Embodiment. It is a figure which shows an example of the content of the evaluation information which concerns on 1st Embodiment. It is a figure which shows an example of the structure of the cell evaluation apparatus which concerns on 1st Embodiment. It is a figure which shows an example of the image to be displayed on the display device when a user is made to select an evaluation condition. It is a figure which shows an example of the effect size of the adhesion cell area index which concerns on 1st Embodiment. It is a figure which shows an example of the code conversion of the index value which concerns on 1st Embodiment. It is a figure which shows an example of the evaluation index group which concerns on 1st Embodiment. It is a flow chart which shows an example of the operation of the cell evaluation apparatus which concerns on 1st Embodiment. It is a flow chart which shows the details of the process of step S130 shown in FIG. It is a figure which shows an example of the image which is displayed when the threshold value of AUC and the threshold value of the evaluation completion time are acquired. It is a figure which shows another example which specifies a suitable time used for evaluation of an evaluation index. It is a figure which shows an example of the image which is displayed when the type of a target cell is acquired. It is a figure which shows an example of the content of the evaluation information at the time of evaluating the non-uniformity of the culture state due to the procedure of the same worker. It is a flow chart which shows an example of the process of evaluating a target cell using another evaluation index group.

[First Embodiment]
Hereinafter, the first embodiment relating to the cell evaluation device and the program will be described with reference to the drawings.

<Outline of cell evaluation system 1>
FIG. 1 is a diagram showing an example of the cell evaluation system 1 according to the first embodiment. As shown in FIG. 1, the cell evaluation system 1 includes a cell evaluation device 10, a storage unit 20, an imaging unit 30, a display device 40, and an input device 50. The storage unit 20, the imaging unit 30, the display device 40, and the input device 50 are connected to the cell evaluation device 10. The cell evaluation system 1 may be provided with an incubator for culturing cells. In this case, the incubator contains a culture container containing the cells to be cultured together with the medium, and the cells are cultured. The cell evaluation system 1 does not have to be provided with an incubator, and cells (culture vessels) to be evaluated by a transport device (not shown) are transported from the incubator outside the cell evaluation system 1 to the cell evaluation system 1 for evaluation. You may.

The imaging unit 30 images the cells (hereinafter, target cell TC) cultured in the medium (the medium shown in the figure) in time series (time-lapse imaging). The imaging unit 30 images the target cell TC and supplies the generated image (hereinafter, target image TP) to the cell evaluation device 10. Hereinafter, the image data representing the signal intensity in each pixel of the image pickup device included in the image pickup unit 30 is also simply referred to as “image”. The storage unit 20 is realized by a RAM (Random Access Memory), an HDD (Hard Disk Drive), a flash memory, or a hybrid storage device in which a plurality of these is combined. Evaluation information 21 is stored in the storage unit 20. The details of the evaluation information 21 will be described later.

The cell evaluation device 10 evaluates the target cell TC shown in the target image TP under various evaluation conditions based on the target image TP and the evaluation information 21. The cell evaluation device 10 outputs an evaluation result image showing the evaluation result of evaluating the target cell TC to the display device 40. The display device 40 displays the evaluation result image output from the cell evaluation device 10. The display device 40 is, for example, a liquid crystal display panel, an organic EL (Electroluminescence) display panel, or the like. The input device 50 accepts the input of information by the user of the cell evaluation device 10 (hereinafter, simply referred to as “user”). The input device 50 is, for example, a keyboard, a mouse, a touch pad, or the like.

FIG. 2 is a diagram showing an example of the contents of the evaluation information 21 according to the first embodiment. The evaluation information 21 includes a type of target cell TC, a period during which the target cell TC was cultured in the medium, a plurality of teacher images PP (illustrated, teacher images PP11 to PP1n) imaged in a time-lapse during the period, and a teacher. The date and time when the image PP was acquired, the evaluation index MK for evaluating the target cell TC, the value of each evaluation index MK calculated based on the associated teacher image PP (hereinafter referred to as the index value MKV), and the culture medium. The procedure (hereinafter referred to as a protocol) and the evaluation results of the respective evaluation conditions CD1 to CD3 are information associated with each identification information (illustrated set ID) that can identify a set of a plurality of teacher image PPs. The teacher image PP is, for example, an image captured before the target image TP is captured. In the following description, a case where the evaluation information 21 includes data acquired for a specific type of cell (“OO cell” in the figure) will be described. Specific types of cells include, for example, somatic cells (nerve cells, blood cells, pancreatic cells, kidney cells, heart cells, etc.), germ cells, stem cells (iPS cells, ES cells, Muse cells, etc.), cancer cells, etc. It is an existing cell.

The evaluation information 21 includes, for example, at least one evaluation index MK among the adhesion cell region index MK1, the floating cell region index MK2, the aggregated cell region index MK3, and the non-cell region index MK4. The adherent cell region index MK1 is an index indicating the area of the region of the monolayer cells adhering to the culture vessel in the teacher image PP. The floating cell region index MK2 is an index showing the area of the region of the cells floating in the medium without the cells adhering to the culture vessel in the teacher image PP. The aggregated cell region index MK3 is an index indicating the area of the region of the multi-layered cell in which a plurality of cells are aggregated, which is adhered to the culture vessel, in the teacher image PP. The non-cell region index MK4 is an index indicating the area of the region where cells do not exist in the teacher image PP.

The protocol is, for example, the type of medium used for culturing, the type of solution added to the medium, the type of reagent added to the medium, the type of dye used for detecting the state of cells, and the sequence of each step of culturing ( For example, the timing of medium replacement).

The evaluation result is the result of final visual evaluation by the user according to the evaluation conditions CD1 to CD3 at the timing when the culture of the target cell TC is completed (for example, good / bad quality, good / bad yield, speed of culture). Is early / late).

When the user visually evaluates the teacher image PP by the evaluation condition CD1 (evaluating the quality of the cells), the cells were obtained in a plurality of series of cultures based on the morphology and color of the similarly cultured cells. The images are sorted into teacher images with good cell quality and teacher images with poor cell quality. In this case, the teacher image PP to be visually evaluated is preferably an image taken in the vicinity of the end time of the culture. For example, when it is evaluated that the cell quality is good for the teacher image taken near the culture end time, the same evaluation may be made for the images acquired in a series of cultures including the teacher image (cell quality is poor). The same applies when evaluated as). The number of teacher image PPs to be visually evaluated in a series of cultures may be one or a plurality. Further, the image does not have to be taken in the vicinity of the end time of the culture. In addition, other existing methods may be used as a method for visually evaluating the quality of cells.

In addition, when the user visually evaluates the teacher image PP according to the evaluation condition CD2 (evaluating the quality of the cell yield), a plurality of series of cultures are performed based on the area of the region where the cultured cells exist. The acquired image is sorted into a teacher image with a good cell yield and a teacher image with a poor cell yield. In this case, the teacher image PP to be visually evaluated is preferably an image taken in the vicinity of the end time of the culture. For example, when the teacher image taken near the culture end time is evaluated as having a good cell yield, the same evaluation may be made for the images acquired in a series of cultures including the teacher image (the cell yield is poor). The same applies when evaluated as). The number of teacher image PPs to be visually evaluated in a series of cultures may be one or a plurality. Further, the image does not have to be taken in the vicinity of the end time of the culture. As a method for visually evaluating the cell yield, another existing method may be used.

In addition, when the user visually evaluates the teacher image PP according to the evaluation condition CD3 (evaluating the quality of the speed at which the target cells are cultured), the number of target cells, the ratio to the total number of cells, etc. Based on this, the images acquired in a plurality of series of cultures are quickly cultured to the target cells (that is, the speed is good) and the teacher images and the target cells are cultured slowly (that is, the speed is good). Select teacher images (which are not fast). Then, when it is evaluated that any teacher image is quickly cultured to the target cells, the same evaluation may be performed for the images acquired in a series of cultures including the teacher image (culture to the target cells). The same applies when the speed is slow). The number of teacher image PPs to be visually evaluated in a series of cultures may be one or a plurality. In addition, other existing methods may be used as a method for visually evaluating the speed at which the target cells are cultured.

The evaluation information 21 may be generated by the cell evaluation device 10, and the evaluation information 21 generated by another device may be stored in the storage unit 20 in advance. Hereinafter, the process when the cell evaluation device 10 generates the evaluation information 21 will be described.

When the cell evaluation device 10 generates the evaluation information 21, the cell evaluation device 10 acquires a plurality of time-lapse imaged teacher image PPs and information related to the teacher image PP from another device. The information related to the teacher image PP is information indicating the date and time when the teacher image PP was imaged and the protocol applied at the timing when the teacher image PP was acquired for each teacher image PP. The cell evaluation device 10 calculates each index value MKV of the adhesion cell region index MK1, the floating cell region index MK2, the aggregated cell region index MK3, and the non-cell region index MK4 for each acquired teacher image PP. Further, the cell evaluation device 10 uses the teacher image PP acquired at the timing when the cell culture shown in the teacher image PP is completed among the acquired plurality of teacher image PPs (that is, the teacher image acquired at the latest date and time). PP) is displayed on the display device 40. The user evaluates the teacher image PP displayed on the display device 40 according to the evaluation conditions CD1 to CD3, and inputs the evaluation result to the input device 50. The cell evaluation device 10 acquires the evaluation result input by the input device 50. Then, the cell evaluation device 10 uses the acquired plurality of teacher image PPs, the date and time when the teacher image PPs are acquired, the index value MKV of each evaluation index MK, the protocol, and the evaluation condition CD at the timing when the culture is completed. The evaluation information 21 is generated and stored in the storage unit 20 by associating the evaluation results with each other.

In the above description, the case where the evaluation result included in the evaluation information 21 is the evaluation result finally evaluated visually by the user at the timing when the culture of the target cell TC is completed has been described, but the present invention is not limited to this. .. For example, the evaluation information 21 may be associated with the evaluation result for each teacher image PP. In this case, the cell evaluation device 10 displays the acquired teacher image PP on the display device 40, respectively. The user evaluates the teacher image PP displayed on the display device 40 according to the evaluation conditions CD1 to CD3, and inputs the evaluation result to the input device 50. The cell evaluation device 10 acquires the evaluation result input by the input device 50. Then, the cell evaluation device 10 includes a plurality of acquired teacher image PPs, a date and time when the teacher image PPs are acquired, an index value MKV of each evaluation index MK for evaluating the target cell TC, a protocol, and an evaluation condition CD. The evaluation information 21 is generated and stored in the storage unit 20 by associating the evaluation results with each other.

Further, in the above description, the case where all the evaluation condition CDs evaluated by the cell evaluation device 10 and the evaluation results of each evaluation condition CD are associated with each teacher image PP of the evaluation information 21 has been described. Not limited. Each teacher image PP of the evaluation information 21 has a configuration in which, for example, a part of the evaluation condition CDs evaluated by the cell evaluation device 10 and the evaluation result of the evaluation condition CD are associated with each other. May be good.

Further, the evaluation information 21 may be associated with the time elapsed since the cell culture was started, instead of (or in addition to) the date and time when the teacher image PP was acquired. Further, the evaluation information 21 does not necessarily have to include the protocol and the evaluation result.

<Overview of cell evaluation device 10>
FIG. 3 is a diagram showing an example of the configuration of the cell evaluation device 10 according to the first embodiment. As shown in FIG. 3, the cell evaluation device 10 includes a control unit 100. The control unit 100 is realized by executing a program (software) by a hardware processor such as a CPU (Central Processing Unit). The control unit 100 includes, for example, a target image acquisition unit 110, an evaluation condition acquisition unit 120, an evaluation index extraction unit 130, an index value calculation unit 140, an evaluation unit 150, and a display control unit 160. Prepare as. In addition, some or all of these components (excluding the contained storage unit) include LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable GPU), and FPGA (Field-Programmable GPU). It may be realized by hardware (circuit unit; including circuit unit) such as Unit), or it may be realized by the cooperation of software and hardware.

The target image acquisition unit 110 acquires the target image TP from the imaging unit 30. The target image acquisition unit 110 supplies the acquired target image TP to the index value calculation unit 140. The evaluation condition acquisition unit 120 acquires information that is input to the input device 50 and indicates at least one of the evaluation conditions CD1 to CD3 selected by the user. FIG. 4 is a diagram showing an example of an image IM1 to be displayed on the display device 40 when the user is made to select the evaluation condition CD. For example, the display device 40 displays the image IM1 that prompts the input of the evaluation condition CD based on the control of the cell evaluation device 10. This image IM1 is, for example, to evaluate the quality of the target cell TC as the evaluation condition CD of the target cell TC (hereinafter, evaluation condition CD1), and to evaluate the quality of the yield of the target cell TC (evaluation condition CD2). ), And the evaluation of the speed at which the target cell TC is cultured to the target cell (hereinafter, evaluation condition CD3), the image that allows the user to select which evaluation condition CD is used to evaluate the target cell TC. Is. The user inputs to the input device 50 an operation of selecting any one of the evaluation condition CDs from the evaluation condition CDs shown in the image IM1.

Returning to FIG. 3, the evaluation condition acquisition unit 120 acquires the evaluation condition CD selected by the user from the input device 50 and supplies it to the evaluation index extraction unit 130. The evaluation index extraction unit 130 extracts the evaluation index MK used when evaluating the evaluation condition CD from the evaluation information 21 based on the evaluation condition CD acquired from the evaluation condition acquisition unit 120 and the evaluation information 21. The evaluation index extraction unit 130 derives a combination of the evaluation indexes MK used for the evaluation of the evaluation unit 150 (hereinafter, evaluation index group MKS) based on the index value MKV of the evaluation index MK extracted from the evaluation information 21. Since the evaluation index extraction unit 130 selects a specific evaluation index MK from among a number of evaluation index MKs, it can be paraphrased as an evaluation index selection unit.

The index value calculation unit 140 is an index of the evaluation index MK included in the evaluation index group MKS based on the target image TP acquired from the target image acquisition unit 110 and the evaluation index group MKS derived by the evaluation index extraction unit 130. The value MKV is calculated based on the target image TP. The evaluation unit 150 evaluates the target cell TC shown in the target image TP according to the evaluation condition CD based on the index value MKV calculated by the index value calculation unit 140. The display control unit 160 causes the display device 40 to display an evaluation result image showing the evaluation result of the evaluation unit 150.

Hereinafter, among the respective units of the control unit 100, the details of the evaluation index extraction unit 130, the index value calculation unit 140, and the evaluation unit 150 will be described.

<About the evaluation index extraction unit 130>
Hereinafter, the details of the processing of the evaluation index extraction unit 130 will be described. First, the evaluation index extraction unit 130 is an evaluation index MK to which the evaluation condition CD acquired from the evaluation condition acquisition unit 120 is associated with the evaluation information 21, and the evaluation is associated with a positive example of the evaluation condition CD. Obtain both the index MK only, the evaluation index MK only with the negative example of the evaluation condition CD, or the evaluation index MK with the positive example of the evaluation condition CD and the evaluation index MK with the negative example associated with it. To do. Hereinafter, it is assumed that the evaluation index extraction unit 130 acquires both the evaluation index MK to which the positive example of the evaluation condition CD is associated and the evaluation index MK to which the negative example is associated. A positive example of the evaluation condition CD is, for example, that the evaluation condition CD1 has "good quality", the evaluation condition CD2 has "good yield", and the evaluation condition CD3 has "fast culture". That is. Negative examples of the evaluation condition CD are, for example, "poor quality" in the case of the evaluation condition CD1, "poor yield" in the case of the evaluation condition CD2, and "slow culture speed" in the case of the evaluation condition CD3. That is.

Next, the evaluation index extraction unit 130 has one or more index values MKV associated with positive examples and one or more evaluation indexes associated with negative examples among the index values MKV extracted from the evaluation information 21. For each MK, a suitable time to be used for evaluation is specified. Hereinafter, a group of one or more evaluation index MKs to which positive examples are associated will be described as a normal case group, and a group of one or more evaluation index MKs to which negative examples are associated will be referred to as a negative example group. The evaluation index extraction unit 130 performs a significant difference test between the extracted positive case group and the negative case group for each evaluation index MK. Hereinafter, a case where the evaluation index extraction unit 130 performs a significant difference test of the adhesion cell region index MK1 among the evaluation indexes MK will be described. The evaluation index extraction unit 130 has a significant probability from the Mann-Whitney U test based on the index value MKV of the adhesion cell region index MK1 of the positive case group and the index value MKV of the adhesion cell region index MK1 of the negative case group. (P value) is calculated. Here, the p-value changes depending on the number of positive case groups, the number of negative case groups, and the total (number of samples). Therefore, the evaluation index extraction unit 130 calculates the z-value from the calculated p-value by the z-test, and calculates the effect size r. The effect size r is a value obtained by dividing the z value by the square root of the number of samples, and is determined by the equation (1). Generally, if the effect size r is 0.5 or more, it can be said that there is a significant difference.

FIG. 5 is a diagram showing an example of the effect size r of the adhesion cell region index MK1 according to the first embodiment. The evaluation index extraction unit 130 sets the time that has passed since the start of culturing the target cell TC and indicates a value (for example, the maximum value) at which the effect size r is 0.5 or more, as the evaluation index MK. It is specified as a suitable time for evaluation. In the example shown in FIG. 5, the effect size r of the adhesion cell region index MK1 shows the largest value at the timing when 246 hours have passed since the culture of the target cell TC was started. Therefore, the evaluation index extraction unit 130 specifies 246 hours after the start of culturing the target cell TC as a suitable time for using the adherent cell region index MK1 for evaluation. The evaluation index extraction unit 130 performs the above-mentioned processing for each evaluation index MK, and specifies a suitable time for using each evaluation index MK for evaluation. The evaluation index extraction unit 130 has specified, but is not limited to, a time indicating a value in which the effect size r is 0.5 or more as a suitable time to be used for the evaluation of the evaluation index MK. For example, the time during which the effect size r is greater than or equal to 0.5 may be specified as a suitable time for evaluation of the evaluation index MK, or the time during which the effect size r is 0.5 or less is evaluated. It may be specified as a suitable time to be used for the evaluation of the index MK.

Next, the evaluation index extraction unit 130 identifies the evaluation index group MKS that can accurately separate the positive example and the negative example of the evaluation condition CD. FIG. 6 is a diagram showing an example of code conversion of the index value MKV according to the first embodiment. Here, it is preferable that the index value MKV changes in the same direction in any of the evaluation indexes MK. Specifically, it is preferable that the index value MKV changes in ascending order from the index value MKV of the negative example group to the index value MKV of the positive example group. Along with this, when the average value of the index value MKV of the positive example group of a certain evaluation index MK is smaller than the average value of the index value MKV of the negative example group, the evaluation index extraction unit 130 reverses the sign of the index value MKV. , The conversion process of converting the value from the index value MKV of the negative example group to the index value MKV of the positive example group in ascending order is performed. If the average value of the index value MKV of the positive example group is larger than the average value of the index value MKV of the negative example group in advance, the evaluation index extraction unit 130 does not need to convert the value.

Since the evaluation index extraction unit 130 only needs to match the direction of change of the index value MKV, for example, for the evaluation index MK in which the average value of the negative example group is smaller than the average value of the positive example group, the index value MKV The sign may be inverted and the value may be converted so as to change from the index value MKV of the positive example group to the index value MKV of the negative example group in ascending order. Further, when the average value of the positive example group of a certain evaluation index MK is smaller than the average value of the negative example group, the evaluation index extraction unit 130 reverses the sign of the index value MKV, but it is not necessary to compare the average values. Good. For example, the evaluation index extraction unit 130 may invert the sign of the index value MKV when the maximum value of the positive example group of a certain evaluation index MK is smaller than the maximum value of the negative example group.

Further, in the present embodiment, the adhesion cell region index MK1, the floating cell region index MK2, the aggregated cell region index MK3, and the non-cell region index MK4 are all evaluation indexes related to the area, but the evaluation of the target cell TC. In some cases, an evaluation index MK that is not an area is used.
In this case, the index value MKV may have different scales. Along with this, the evaluation index extraction unit 130 performs standardization processing for aligning the scale of each index value MKV. Specifically, the evaluation index extraction unit 130 performs standardization processing in which the average of each index value MKV is "0" and the variance is "1". Specifically, each index value MKV is converted by the equation (2).

The evaluation index extraction unit 130 weights the index value MKV based on the index value MKV subjected to the above-mentioned code conversion processing and the effect size r. Here, the effect size r takes any value from "0" to "1", and the closer the value is to "1", the more significant the difference. The evaluation index extraction unit 130 weights the index value MKV by the equation (3).

The evaluation index extraction unit 130 identifies the evaluation index group MKS that can accurately separate the positive example and the negative example of the evaluation condition CD based on the index value MKV after weighting. The evaluation index extraction unit 130 derives a ROC (Receiver Operating Characteristic) curve from the positive case group and the negative case group for the index value MKV, and among all the combinations of the evaluation index MK, the positive example of the evaluation condition CD The evaluation index group MKS whose AUC (Area Under the Curve), which is a value indicating the accuracy of separating the negative example, is close to "1" (that is, the separation accuracy between the positive example and the negative example is high) is specified. Depending on the value of AUC for the combination of the evaluation index MK, the evaluation index MK included in the evaluation index group MKS may be one, but even if the evaluation index MK is one, it is referred to as the evaluation index group MKS. Since the method of deriving the ROC curve and the method of calculating the AUC are known, the description of the method will be omitted. The evaluation index extraction unit 130 does not have to specify the evaluation index group MKS whose AUC value is closest to "1".

FIG. 7 is a diagram showing an example of the evaluation index group MKS according to the first embodiment. FIG. 7 is a diagram showing an example of the evaluation index group MKS used when evaluating the target cell TC under the evaluation condition CD1. In the example shown in FIG. 7, the evaluation index group MKS having the highest separation accuracy (that is, the value of AUC is high (AUC = 0.957 in the figure)) is 162 after the start of culturing the target cell TC. The floating cell region index MK2 after an hour, the aggregated cell region index MK3 162 hours after the start of culturing the target cell TC, and the adherent cell region index MK1 147 hours after the start of culturing the target cell TC. And are included.

Here, the value representing the accuracy of separating the positive and negative examples of the evaluation condition CD can be rephrased as the value representing the accuracy of separating the quality of the evaluation condition CD. Therefore, the value representing the accuracy of separating the quality of the evaluation condition CD is not limited to AUC. For example, the above-mentioned effect size r is also a value indicating the accuracy of separating the quality of the evaluation condition CD, and other known methods for obtaining the accuracy of separating the positive example and the negative example may be used. The accuracy of separating positive and negative cases is such that the evaluation target (cell quality in the case of evaluation condition CD1) can accurately separate positive and negative cases, and the higher the accuracy, the higher the accuracy. The more accurately the target to be evaluated can be separated into positive and negative cases.

The evaluation index extraction unit 130 performs a process of associating the separation accuracy value, the evaluation index group MKS, and the evaluation condition CD in advance based on the evaluation information 21, and stores the evaluation information 21 generated by the process. It may be stored in the part 20. Further, the weighting of the evaluation index MK by the effect size r does not necessarily have to be performed.

<About index value calculation unit 140>
Hereinafter, the details of the processing of the index value calculation unit 140 will be described. First, the index value calculation unit 140 calculates the index value MKV by using, for example, the evaluation index group MKS having the highest separation accuracy among the evaluation index group MKS associated with the evaluation index extraction unit 130. For example, when the index value MKV is calculated based on the evaluation index group MKS having the highest separation accuracy among the evaluation index group MKS shown in FIG. 7, the index value calculation unit 140 takes 162 hours from the start of culturing the target cell TC. Later, the index value MKV of the floating cell region index MK2 based on the target image TP, the index value MKV of the aggregated cell region index MK3 based on the target image TP 162 hours after the start of culturing the target cell TC, and the culture of the target cell TC. The adherent cell region index MK1 based on the target image TP 147 hours after the start is calculated. The index value MKV calculated by the index value calculation unit 140 is subjected to the above-mentioned code conversion processing and weighting. When the scales of the index value MKV are different, the index value calculation unit 140 may perform the standardization process described above.

The timing at which the index value calculation unit 140 calculates the index value MKV is, for example, the same timing as the timing at which the target image TP is acquired from the target image acquisition unit 110. The timing at which the index value calculation unit 140 calculates the index value MKV may be different from the timing at which the target image TP is acquired. As an example, the index value calculation unit 140 may calculate the index value MKV at predetermined time intervals after acquiring the target image TP. The predetermined time may be, for example, a time interval in which changes in cells can be detected (for example, every few hours). Further, the index value calculation unit 140 calculates the total sum of each index value MKV (hereinafter, total index value MKVT) when there are a plurality of evaluation index MKs included in the evaluation index group MKS. When there is one evaluation index MK included in the evaluation index group MKS, the index value calculation unit 140 calculates the index value MKV of the evaluation index MK as the total index value MKVT. The evaluation index extraction unit 130 may supply statistical values such as an average value and a median value of a plurality of index values MKV to the evaluation unit 150 instead of each total index value MKVT.

<About evaluation unit 150>
Hereinafter, the details of the processing of the evaluation unit 150 will be described. The evaluation unit 150 evaluates the target cell TC shown in the target image TP according to the evaluation condition CD input by the user, based on the total index value MKVT calculated by the index value calculation unit 140. The evaluation unit 150 evaluates the target cell TC using, for example, a value obtained by multiplying the total index value MKVT calculated by the index value calculation unit 140 by a predetermined value as a threshold value. When the threshold value to be evaluated is different for each evaluation index MK, the value obtained by multiplying the total index value MKVT by a predetermined value may be the sum of the threshold values for each evaluation index MK, or the evaluation index MK. It may be a value obtained by multiplying the average value of each threshold value by the number of evaluation indexes MK included in the evaluation index group MKS. For example, among the evaluation index group MKS shown in FIG. 7, the evaluation index group MKS having the highest separation accuracy includes three evaluation index MKs. In this case, the evaluation unit 150 evaluates the target cell TC by comparing the threshold value, which is a value obtained by multiplying a predetermined value by "3", with the total index value MKVT calculated by the index value calculation unit 140. .. For example, when the total sum of the total index values MKVT is equal to or greater than the threshold value, the evaluation unit 150 evaluates the target cell TC as a state in which a positive example of the evaluation condition CD is satisfied. Further, the evaluation unit 150 evaluates the target cell TC as a state in which a negative example of the evaluation condition CD is satisfied when the total index value MKVT is less than the threshold value. The evaluation unit 150 supplies information indicating the evaluation result to the display control unit 160. When the total index value MKVT acquired from the evaluation index extraction unit 130 is the statistical value (for example, the average value or the median value) of the index value MKV, the evaluation unit 150 uses the threshold value of the statistical value to target. Cell TC shall be evaluated.

<About the operation of the cell evaluation device 10>
FIG. 8 is a flow chart showing an example of the operation of the cell evaluation device 10 according to the first embodiment. The target image acquisition unit 110 acquires the target image TP captured by the image pickup unit 30. (Step S110). Next, the evaluation condition acquisition unit 120 acquires the evaluation condition CD input to the input device 50 (step S120). Next, the evaluation index extraction unit 130 derives the evaluation index group MKS based on the acquired evaluation condition CD and the evaluation information 21 (step S130). Details of the process in step S130 will be described later. Next, the index value calculation unit 140 calculates the total index value MKVT of the target cell TC shown in the target image TP based on the evaluation index group MKS derived by the evaluation index extraction unit 130 (step S140). For example, the index value calculation unit 140 weights the evaluation index group MKS derived by the evaluation index extraction unit 130 with respect to the effect size r, and calculates the total index value MKVT. Next, the evaluation unit 150 evaluates the target cell TC shown in the target image TP according to the evaluation condition CD based on the total index value MKVT calculated by the index value calculation unit 140 (step S150). When the total index value MKVT is equal to or higher than the threshold value (step S150; YES), the evaluation unit 150 evaluates (determines) that the target cell TC is in a state in which a positive example of the evaluation condition CD is satisfied (step S160). .. Further, the evaluation unit 150 evaluates that the target cell TC is in a state where a negative example of the evaluation condition CD is satisfied when the index value MKV or each total index value MKVT is less than the threshold value (step S150; NO) (step S150; NO). (Determination) (step S170). The display control unit 160 displays an evaluation result image showing the evaluation result evaluated by the evaluation unit 150 on the display device 40 (step S180).

FIG. 9 is a flow chart showing details of the process of step S130 shown in FIG. First, the evaluation index extraction unit 130 extracts the index value MKV of the normal example group related to the evaluation condition CD from the evaluation information 21 based on the evaluation condition CD acquired by the evaluation condition acquisition unit 120 (step S200). .. Next, the evaluation index extraction unit 130 extracts the index value MKV of the negative example group related to the evaluation condition CD from the evaluation information 21 (step S202). The evaluation index extraction unit 130 calculates the effect size r based on the index value MKV of the positive example group and the index value MKV of the negative example group (step S204). Next, the evaluation index extraction unit 130 performs code change processing on the index value MKV so as to change from the index value MKV of the negative example group to the index value MKV of the positive example group in ascending order (step S206). Next, the evaluation index extraction unit 130 weights the index value MKV after the code conversion process based on the effect size r (step S208). Next, the evaluation index extraction unit 130 has an AUC close to "1", which is a value indicating the accuracy of separating the positive example and the negative example of the evaluation condition CD based on the index value MKV after weighting. (That is, the evaluation index group MKS (that is, the separation accuracy between the positive example and the negative example is high) is specified (step S210).

If the index values MKV acquired in steps S200 to S202 are already in ascending order from the index value MKV of the negative example group to the index value MKV of the positive example group, the process of step S206 does not have to be performed. When the scales of the index value MKV are different, a standardization process for standardizing the index value MKV may be performed instead of (or in addition to) the process of step S206.

Further, in the above description, the case where the evaluation condition acquisition unit 120 acquires any one of the evaluation conditions CD1 to CD3 has been described, but the present invention is not limited to this. For example, the evaluation condition acquisition unit 120 may acquire a plurality of evaluation condition CDs among the evaluation conditions CD1 to CD3. In this case, the user refers to the image IM1 (see FIG. 4) displayed on the display device 40, and inputs an operation of selecting a plurality of evaluation condition CDs to the input device 50.
In the above description, after the evaluation condition acquisition unit 120 acquires the evaluation condition CD (step S120), the evaluation index extraction unit 130 derives the evaluation index group MKS (step S130), but the evaluation index extraction unit 130 has previously derived the evaluation index group MKS. The evaluation index group MKS may be derived. In this case, for example, the evaluation index extraction unit 130 derives the evaluation index group MKS for each evaluation condition CD using a series of target image TPs (teacher image PP) acquired in advance by the imaging unit 30, and stores the evaluation index group MKS in the storage unit 20. Remember it. Then, after the evaluation condition acquisition unit 120 acquires the evaluation condition CD (step S120), in step S140, the index value calculation unit 140 reads out the evaluation index group MKS corresponding to the acquired evaluation condition from the storage unit 20, and the index value calculation unit 140 reads the evaluation index group MKS. The total index value MKVT of the evaluation index group MKS may be calculated. In other words, as the operation of the cell evaluation device 10, step S130 (subroutine in FIG. 9) of FIG. 8 is executed in advance, and then the operation of steps S110 to S180 excluding step S130 is executed as described above. To.

Further, in the above description, the case where the cell evaluation device 10 acquires both the evaluation index group MKS and the suitable time of the index value MKV has been described, but the present invention is not limited to this. The evaluation index extraction unit 130 may, for example, derive only the evaluation index group MKS, or may calculate only a suitable time of the index value MKV. When only the evaluation index group MKS is used, the evaluation index extraction unit 130 calculates the average value of the AUC values at each time for each evaluation index MK. Then, the evaluation index extraction unit 130 compares the average value of each index value MKV with the threshold value, and derives the evaluation index group MKS that is equal to or greater than the threshold value. When calculating only the suitable time of the index value MKV, the evaluation index extraction unit 130 calculates the average value of the effect size r at each time for each evaluation index MK. Then, the evaluation index extraction unit 130 compares the average value of each index value MKV with the threshold value, and derives the evaluation index group MKS that is equal to or greater than the threshold value. The evaluation index extraction unit 130 may derive the evaluation index group MKS using other statistical values instead of the average value.

<When specifying the ACU value and evaluation completion time>
In the above description, the case where only the evaluation condition CD is accepted by the input device 50 has been described, but the present invention is not limited to this. The input device 50 may be configured to receive information other than the evaluation condition CD as long as it is information related to evaluation, for example. FIG. 10 is a diagram showing an example of the image IM2 displayed when acquiring the threshold value of AUC and the threshold value of the evaluation completion time. As shown in FIG. 10, the display control unit 160 performs the time from the start of culturing the AUC threshold value and the target cell TC on the display device 40 to the completion of the evaluation by the evaluation condition CD (hereinafter, the evaluation completion time). ) Is displayed as the image IM2 that accepts the threshold value. The AUC threshold is input to the input device 50 by the user. In this case, the evaluation index extraction unit 130 identifies the evaluation index group MKS that is equal to or higher than the AUC threshold value received by the input device 50. As a result, the cell evaluation device 10 can evaluate the target cell TC by the evaluation index group MKS in which the separation accuracy desired by the user is maintained. Further, the threshold value of the evaluation completion time is input to the input device 50 by the user. In this case, the evaluation index extraction unit 130 specifies the evaluation index group MKS in which the time when the effect size r is 0.5 or more is the time until the evaluation completion time received by the input device 50. As a result, the cell evaluation device 10 can evaluate the target cell TC by the user's desired evaluation completion time. It is not necessary to display the AUC value on the display device 40. For example, the display device 40 may be divided into three, high, medium, and low, depending on the magnitude of the AUC value, and any of the selections may be accepted. Further, the image IM2 may be an image that accepts either the AUC threshold value or the evaluation completion time. Further, the vertical axis of the image IM2 may be the effect size r instead of the AUC.

<Another example of a method for specifying a suitable time using the evaluation index MK for evaluation>
Further, in the above description, the evaluation index extraction unit 130 uses the time indicating a value (for example, the maximum value) at which the effect size r is 0.5 or more after the start of culturing the target cell TC as the evaluation of the evaluation index MK. The case of specifying as a suitable time to be used has been described, but the present invention is not limited to this. The evaluation index extraction unit 130 may be configured to specify a suitable time to be used for the evaluation of the evaluation index MK, for example, based on a value other than the effect size r.

FIG. 11 is a diagram showing another example of specifying a suitable time used for evaluation of the evaluation index MK. Specifically, FIG. 11 shows the difference between the average value of the positive example group of the index value MKV of a certain evaluation index MK and the average value of the negative case group, and the average of the positive example group of the index value MKV of a certain evaluation index MK. It is a figure which shows the time change of the ratio of the value, and the average value of a negative example group. The evaluation index extraction unit 130 evaluates, for example, the time when the difference between the average value of the positive example group and the average value of the negative case group of the index value MKV of a certain evaluation index MK is the largest (that is, the time when there is a significant difference). It may be specified as a suitable time to be used for the evaluation of the index MK. Further, in the evaluation index extraction unit 130, for example, the time when the ratio of the average value of the positive example group of the index value MKV of a certain evaluation index MK to the average value of the negative case group is the largest (that is, the time when there is a significant difference). May be specified as a suitable time to be used for the evaluation of the evaluation index MK.

The evaluation index extraction unit 130 specifies a suitable time to be used for the evaluation of the evaluation index MK only by the difference between the average value of the positive example group and the average value of the negative example group of the index value MKV of a certain evaluation index MK. Alternatively, the suitable time to be used for the evaluation of the evaluation index MK may be specified only by the ratio of the average value of the positive example group and the average value of the negative example group of the index value MKV of a certain evaluation index MK. , The suitable time used to evaluate the evaluation index MK may be specified based on both the difference and the ratio values. The evaluation index extraction unit 130 evaluates the evaluation index MK when the difference between the average value of the positive example group and the average value of the negative example group of the index value MKV of a certain evaluation index MK is larger than a predetermined value. It may be specified as a suitable time to be used, and the time when the ratio of the average value of the positive example group and the average value of the negative example group of the index value MKV of a certain evaluation index MK is larger than a predetermined value is specified as the evaluation index MK. It may be specified as a suitable time to be used for the evaluation of. Further, the evaluation index extraction unit 130 sets the effect amount r of the normal case group based on the index value MKV of the positive case group of a certain evaluation index MK and the effect amount r of the negative case group based on the index value MKV of the negative case group. The time calculated for each time and the difference between the effect amount r of the positive case group and the effect amount r of the negative case group is the largest may be specified as a suitable time to be used for the evaluation of the evaluation index MK. The evaluation index extraction unit 130 specifies a time in which the difference between the effect size r of the positive case group and the effect size r of the negative case group is larger than a predetermined value as a suitable time to be used for the evaluation of the evaluation index MK. Alternatively, a time in which the ratio of the effect size r of the positive case group to the effect size r of the negative case group is larger than a predetermined value may be specified as a suitable time to be used for the evaluation of the evaluation index MK. ..

[Second Embodiment]
Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. In the second embodiment, a case where cells are evaluated by evaluation condition CDs other than evaluation conditions CD1 to CD3 will be described. The same components as those in the first embodiment described above are designated by the same reference numerals, and the description thereof will be omitted.

<About cell mistake evaluation>
Hereinafter, a case where the cell evaluation device 10 evaluates the mistake of the target cell TC will be described with reference to the drawings. In the following description, evaluating the mistake of the target cell TC will be referred to as "evaluation condition CD4". FIG. 12 is a diagram showing an example of an image IM3 displayed when acquiring the type of target cell TC. Here, cells different from the target to be cultured (hereinafter referred to as foreign cells) are cultured as the target cell TC due to the fact that incorrect cells are collected as the target cell TC from the sample or the medium is mistaken due to human error. May occur. In this case, it is preferable that the cell evaluation device 10 evaluates the target cell TC and presents to the user that the foreign cells are cultured as the target cell TC. The details of the cell evaluation device 10 when performing this evaluation will be described. When the cell evaluation device 10 evaluates the target cell TC according to the evaluation condition CD4, the evaluation information 21 includes data (teacher) acquired for a plurality of types of cells (for example, “□□ cells” and “△△ cells”). Image PP, evaluation index MK, index value MKV, protocol, evaluation result of evaluation condition CD4, etc.) are included. The configuration for generating the evaluation information 21 for a plurality of types of cells other than the “XX cells” is the same as that of the first embodiment described above, and thus the description thereof will be omitted.

The user refers to the image IM3 displayed on the display device 40, and inputs the type of the target cell TC (that is, a cell that is not a foreign cell) that the user wants to culture in the input device 50. In the illustrated example, the user inputs "XX cells" as the type of target cell TC via the input device 50. The evaluation index extraction unit 130 extracts the evaluation index MK of the positive example group of the evaluation condition CD4 and the evaluation index MK of the negative example group of the evaluation condition CD4 based on the evaluation information 21, and derives the evaluation index group MKS. .. At this time, the evaluation index extraction unit 130 targets the evaluation index group MKS having high separation accuracy between the positive example and the negative example, and the appropriate time for using the index value MKV included in the evaluation index group MKS for evaluation. An evaluation index group MKS, which is an earlier time after the start of cell TC culture, is derived. After that, the index value calculation unit 140 calculates the total index value MKVT based on the evaluation index group MKS derived by the evaluation index extraction unit 130 and the target image TP, and the evaluation unit 150 calculates the index value. Since the process of evaluating the target cell TC under the evaluation condition CD4 based on the index value MKV calculated by the unit 140 is the same as described above, the description thereof will be omitted.

The display control unit 160 causes the display device 40 to display an image showing the evaluation result of the evaluation unit 150. For example, when the evaluation result indicates that the foreign cells are cultured as the target cell TC, the display control unit 160 is presented with, for example, "a cell different from the intended cell is cultured". The image is displayed on the display device 40. For example, when the target image TP is a target image TP related to the target cell TC in the middle of culturing, the user can use the target cell TC (that is, a foreign cell) in which the culturing cell is desired to be cultured based on the display. Since it is possible to know whether or not the cells are not cells), unnecessary culture can be stopped immediately.

Note that the user may be configured to set a threshold value for the evaluation completion time, for example. In this case, the display control unit 160 displays the image IM that receives the threshold value of the evaluation completion time on the display device 40, and acquires the threshold value by the input device 50. The evaluation index extraction unit 130 derives an evaluation index group MKS including an index value MKV whose suitable time used for evaluation is equal to or less than the threshold value. Thereby, by the time desired by the user, it is possible to present whether or not the cell being cultured is the target cell TC that the user desires to culture. The threshold value of the evaluation completion time may be automatically set by the cell evaluation device 10. In this case, the cell evaluation device 10 may acquire and set a value used with high frequency as a threshold value of the evaluation completion time of the target cell TC from the past log, and the evaluation of the target cell TC is completed. It may set an adaptable minimum value as a time threshold.

The evaluation index extraction unit 130 includes the evaluation index group MKS in which the index value MKV of the cell intended by the user is a positive example group and the index value MKV of another cell is a negative example group in the evaluation information 21. It may be derived for each type of. The evaluation unit 150 can identify which cell the foreign cell is, for example, based on each total index value MKVT calculated based on the evaluation index group MKS for each cell. In this case, the display control unit 160 displays an image showing the type of different cells such as "▲▲ cells different from the intended cells are cultured" on the display device 40, and displays a more detailed evaluation result by the user. Can be presented to.

<Evaluation of non-uniformity in culture state>
Hereinafter, a case where the cell evaluation device 10 evaluates the non-uniformity of the culture state of the target cell TC will be described. Here, when a procedure is performed by an operator in a certain protocol, the state of culturing the target cell TC may be uneven due to the variation in the procedure. Non-uniform culture of the target cell TC means, for example, that the speed of differentiation varies among a plurality of cells, the speed of maturation varies among a plurality of cells, or the culture vessel (medium). The amount of contamination generated in the cell varies. In this case, it is preferable that the cell evaluation device 10 evaluates the heterogeneity of the target cell TC with respect to the procedure performed in the culture of the target cell TC. Hereinafter, the details of the cell evaluation device 10 when performing this evaluation will be described. Hereinafter, the cell evaluation device 10 will be described assuming that only cells of the same type are cultured.

<Evaluation of non-uniformity of culture state due to the procedure of the same worker>
Hereinafter, a case where the cell evaluation device 10 evaluates the non-uniformity of the procedure of the same operator will be described. FIG. 13 is a diagram showing an example of the contents of the evaluation information 21 when evaluating the non-uniformity of the culture state due to the procedure of the same operator. In the following description, evaluating the non-uniformity of the culture state due to the procedure of the same operator will be referred to as "evaluation condition CD8". As shown in FIG. 13, when the cell evaluation device 10 evaluates the target cell TC according to the evaluation condition CD8, the evaluation information 21 contains information that can identify the worker who performed the protocol (hereinafter, worker ID). It is associated with each protocol. The worker ID may be, for example, associated with the teacher image PP set by a known method by the cell evaluation device 10 based on the work history information (not shown) of the worker, and the procedure is performed. The worker ID is assigned by arranging an object indicating his / her own worker ID in the vicinity of the medium so that the image pickup unit 30 can image the worker ID, and the cell evaluation device 10 recognizes the teacher image PP as an image. It may be the one to be acquired. Further, in the evaluation information 21, the evaluation result of evaluating the teacher image PP according to the evaluation condition CD8 (evaluating the presence or absence of non-uniformity of the culture state due to the procedure of the same worker) is the evaluation result of the plurality of teacher image PPs. It is associated with each set. Hereinafter, a specific configuration of the cell evaluation device 10 that generates the evaluation information 21 in this case will be described.

The cell evaluation device 10 is a teacher image PP set associated with the same worker ID, and is a teacher image PP set relating to the first cell and the second cell cultured in the same well and in different media. To get each. The cell evaluation device 10 displays the latest two teacher image PPs acquired at the same date and time on the display device 40 side by side among the acquired first cell and the plurality of teacher image PPs related to the second cells. The user evaluates the two teacher image PPs displayed on the display device 40 according to the evaluation condition CD8, and inputs the evaluation result to the input device 50. Specifically, when the user has non-uniformity in the two teacher image PPs displayed on the display device 40 (for example, the first cell is in a good culture state and the second cell is in a poor culture state). Etc.), for the teacher image PP set related to the first cell, input that there is no non-uniformity as the evaluation result of the evaluation condition CD8 to the input device 50, and for the teacher image PP set related to the second cell. Inputs to the input device 50 that there is non-uniformity as the evaluation result of the evaluation condition CD8. The cell evaluation device 10 acquires the evaluation result input by the input device 50. Then, the cell evaluation device 10 evaluates the acquired plurality of teacher image PPs, the date and time when the teacher image PPs are acquired, the index value MKV of each evaluation index MK, the protocol, the worker ID, and the evaluation condition CD8. The evaluation information 21 is generated by associating the results with each other and stored in the storage unit 20.

The evaluation index extraction unit 130 is an evaluation index MK associated with the same worker ID as the worker who performed the protocol of the target cell TC based on the evaluation information 21, and evaluates the normal group of the evaluation condition CD8. The index MK and the evaluation index MK of the negative case group of the evaluation condition CD8 are acquired. After that, the evaluation index extraction unit 130 derives the evaluation index group MKS based on the evaluation index MK of the positive case group and the evaluation index MK of the negative case group, and the index value calculation unit 140 extracts the evaluation index. The process of calculating the total index value MKVT based on the evaluation index group MKS derived by the unit 130 and the target image TP, and the evaluation unit 150 based on the index value MKV calculated by the index value calculation unit 140. Since the process of evaluating the target cell TC according to the evaluation condition CD8 is the same as described above, the description thereof will be omitted.

When the evaluation result of the evaluation unit 150 indicates that the target cell TC is non-uniform due to the procedure by the same operator, the display control unit 160 displays an image indicating that the non-uniformity is occurring. Display on 40. For example, when the target image TP is the target image TP related to the target cell TC in the middle of culturing, the user can know that non-uniformity has occurred during culturing based on the display. Non-uniformity can be eliminated in real time by adding additives, changing the medium, etc.). Further, when the target image TP is the target image TP related to the target cell TC in the middle of culturing, the user can use the time zone in which the non-uniformity occurs (that is, the procedure that causes the non-uniformity) based on the display. Since it is possible to grasp the above, it is possible to give feedback to the operator (for example, training of the operator's procedure or alerting the operator) to eliminate the variation in the procedure. Further, by eliminating the non-uniformity by these treatments, the quality and yield of the target cell TC can be improved.

<Evaluation of non-uniformity of culture state due to different operator's procedure>
Hereinafter, a case of evaluating the non-uniformity of the culture state due to the procedure of different workers will be described. In the following description, evaluating the non-uniformity of the culture state due to the procedure of different workers will be referred to as "evaluation condition CD9". Further, in this case, the evaluation information 21 is associated with the evaluation result of evaluating the teacher image PP according to the evaluation condition CD9 for each set of the plurality of teacher image PPs. Hereinafter, a specific configuration of the cell evaluation device 10 that generates the evaluation information 21 in this case will be described.

The cell evaluation device 10 is a teacher image PP set to which the same protocol is associated, a teacher image PP set to which a certain worker ID (hereinafter, worker ID 1) is associated, and a worker other than the worker ID 1. A teacher image PP set associated with an ID (hereinafter, worker ID 2) is acquired. The cell evaluation device 10 displays the latest and two teacher image PPs acquired at the same date and time side by side on the display device 40 among the plurality of teacher image PPs related to the acquired worker ID1 and the worker ID2. The user evaluates the two teacher image PPs displayed on the display device 40 according to the evaluation condition CD9, and inputs the evaluation result to the input device 50. Specifically, when the user has non-uniformity in the two teacher image PPs displayed on the display device 40 (for example, the cells of the worker ID 1 are in a good culture state, and the cells of the worker ID 2 are in a good culture state. (The culture state is poor, etc.), for the teacher image PP set related to the worker ID1, input to the input device 50 that there is no non-uniformity as the evaluation result of the evaluation condition CD9, and the teacher image PP related to the worker ID2. For the set, it is input to the input device 50 that there is non-uniformity as the evaluation result of the evaluation condition CD9. The cell evaluation device 10 acquires the evaluation result input by the input device 50. Then, the cell evaluation device 10 evaluates the acquired plurality of teacher image PPs, the date and time when the teacher image PPs are acquired, the index value MKV of each evaluation index MK, the protocol, the worker ID, and the evaluation condition CD9. The evaluation information 21 is generated by associating the results with each other and stored in the storage unit 20.

The evaluation index extraction unit 130 is an evaluation index MK associated with the same worker ID as the worker who performed the protocol of the target cell TC based on the evaluation information 21, and evaluates the normal group of the evaluation condition CD9. The index MK and the evaluation index MK of the negative case group of the evaluation condition CD9 are acquired. After that, the evaluation index extraction unit 130 derives the evaluation index group MKS based on the evaluation index MK of the positive case group and the evaluation index MK of the negative case group, and the index value calculation unit 140 extracts the evaluation index. The process of calculating the total index value MKVT based on the evaluation index group MKS derived by the unit 130 and the target image TP, and the evaluation unit 150 based on the index value MKV calculated by the index value calculation unit 140. Since the process of evaluating the target cell TC according to the evaluation condition CD9 is the same as described above, the description thereof will be omitted.

When the evaluation result of the evaluation unit 150 indicates that the target cell TC is non-uniform due to a procedure performed by a different operator, the display control unit 160 displays an image indicating that the non-uniformity is occurring. Display on 40. For example, when the target image TP is the target image TP related to the target cell TC in the middle of culturing, the user can know that non-uniformity has occurred during culturing based on the display. By doing so, non-uniformity can be eliminated in real time. In addition, since the user can grasp that non-uniformity has occurred for a certain worker based on the display, feedback to the worker for eliminating the variation in the procedure (for example, the worker). (Training of procedures, alerting workers, etc.) can be performed. Further, by eliminating the non-uniformity by these treatments, the quality and yield of the target cell TC can be improved.

<Evaluation of non-uniformity of culture state due to procedures between different experiments>
Hereinafter, a case of evaluating the non-uniformity of the culture state due to the procedure between different experiments will be described. In the following description, evaluating the non-uniformity of the culture state due to the procedure between different experiments will be referred to as "evaluation condition CD10". Further, in this case, the evaluation information 21 is associated with the evaluation result of evaluating the teacher image PP according to the evaluation condition CD10 for each set of the plurality of teacher image PPs. Hereinafter, a specific configuration of the cell evaluation device 10 that generates the evaluation information 21 in this case will be described.

The cell evaluation device 10 is a teacher image PP set associated with the same worker ID, and acquires a teacher image PP set associated with different protocols. Hereinafter, it is assumed that the different protocols are Protocol A and Protocol B. The cell evaluation device 10 displays two teacher image PPs acquired at the same date and time on the display device 40 side by side among the acquired teacher image PPs of protocol A and protocol B. The user evaluates the two teacher image PPs displayed on the display device 40 according to the evaluation condition CD10, and inputs the evaluation result to the input device 50. Specifically, when the user has non-uniformity in the two teacher image PPs displayed on the display device 40 (for example, the cells of protocol A are in a good culture state, and the cells of protocol B are in a culture state). For the teacher image PP set related to protocol A, input that there is no non-uniformity as the evaluation result of the evaluation condition CD10 to the input device 50, and for the teacher image PP set related to protocol B. Inputs to the input device 50 that there is non-uniformity as an evaluation result of the evaluation condition CD10. The cell evaluation device 10 acquires the evaluation result input by the input device 50. Then, the cell evaluation device 10 evaluates the acquired plurality of teacher image PPs, the date and time when the teacher image PPs are acquired, the index value MKV of each evaluation index MK, the protocol, the worker ID, and the evaluation condition CD10. The evaluation information 21 is generated by associating the results with each other and stored in the storage unit 20.

Based on the evaluation information 21, the evaluation index extraction unit 130 is associated with the same worker ID as the worker who performed the protocol of the target cell TC, and the evaluation index MK associated with a protocol different from the target cell TC. Therefore, the evaluation index MK of the positive case group of the evaluation condition CD10 and the evaluation index MK of the negative case group of the evaluation condition CD10 are acquired. After that, the evaluation index extraction unit 130 derives the evaluation index group MKS based on the evaluation index MK of the positive case group and the evaluation index MK of the negative case group, and the index value calculation unit 140 extracts the evaluation index. The process of calculating the total index value MKVT based on the evaluation index group MKS derived by the unit 130 and the target image TP, and the evaluation unit 150 based on the index value MKV calculated by the index value calculation unit 140. Since the process of evaluating the target cell TC according to the evaluation condition CD10 is the same as described above, the description thereof will be omitted.

When the evaluation result of the evaluation unit 150 indicates that the target cell TC has non-uniformity due to a procedure between different experiments, the display control unit 160 displays an image indicating that the non-uniformity has occurred. Display on 40. For example, when the target image TP is the target image TP related to the target cell TC in the middle of culturing, the user can know that non-uniformity has occurred during culturing based on the display. By doing so, non-uniformity can be eliminated in real time. In addition, since the user can grasp that non-uniformity has occurred in a certain protocol based on the display, feedback to the operator for eliminating the variation in the procedure (for example, the operator's) Procedure training, alerting workers, etc.) can be performed. Further, by eliminating the non-uniformity by these treatments, the quality and yield of the target cell TC can be improved.

<About evaluation results>
In the above description, the case where the evaluation result included in the evaluation information 21 is represented by two values (for example, good / bad, fast / slow, yes / non, etc.) has been described, but the evaluation result has two or more values (for example, good / bad). , Gradual, or linear). Hereinafter, a specific configuration of the cell evaluation device 10 that generates the evaluation information 21 in this case will be described. The cell evaluation device 10 causes the display device 40 to display the teacher image PP included in the teacher image PP set. When the user visually evaluates the teacher image PP displayed on the display device 40 by each evaluation condition CD, the user satisfies the evaluation condition CD in the target cell TC included in the teacher image PP (that is, a positive example). ) The ratio of the target cell TC and the ratio of the target cell TC that does not satisfy the evaluation condition CD (that is, a negative example) are input to the input device 50 as the evaluation result. Then, the cell evaluation device 10 generates evaluation information 21 in which the ratio input as the evaluation result is associated with the teacher image PP or the like. By evaluating using the evaluation information 21, the cell evaluation device 10 can present the evaluation result evaluated by each evaluation condition CD to the user in more detail. By confirming the evaluation result, the user can immediately stop unnecessary culture if the positive case rate of the evaluation result of the target cell TC is too low (or the negative case rate is too high). In addition, by confirming the evaluation result, the user adds when the negative case rate of the evaluation result of the target cell TC is low (or the positive case rate is high), as compared with the case where the evaluation result is binary. It becomes possible to consider taking measures such as adding substances and exchanging the medium, and it is possible to determine the guideline for culturing the target cell TC.

<About change of evaluation index group MKS>
Further, in the above description, the case where the same evaluation index group MKS is used for any target image TP captured in time-lapse has been described, but the present invention is not limited to this. For example, the cell evaluation device 10 evaluates the target cell TC by using another evaluation index group MKS when the desired evaluation result is not obtained as a result of evaluating the target cell TC based on a certain evaluation index group MKS. It may be something to do. FIG. 14 is a flow chart showing an example of processing for evaluating the target cell TC using another evaluation index group MKS. Since the processes from step S110 to step S180 shown in FIG. 14 are the same as the processes of step S having the same reference numerals as shown in FIG. 8, the description thereof will be omitted. In this example, the user asks "whether or not to further evaluate using the evaluation index group MKS other than the evaluation index group MKS used for this evaluation" on the evaluation result image displayed on the display device 40 in step S180. The image to be inquired about shall be included.

The cell evaluation device 10 is an input indicating that after the evaluation result image is displayed on the display device 40 in step S180, further evaluation is performed using the evaluation index group MKS other than the evaluation index group MKS used for this evaluation. Is determined whether or not the input device 50 has been performed (step S190). The cell evaluation device 10 ends the process when no input indicating that further evaluation is performed is performed on the input device 50. When the input device 50 is input to indicate that further evaluation is to be performed, the evaluation index extraction unit 130 has the evaluation index group MKS having the next highest separation accuracy after the evaluation index group MKS used in the previous evaluation. (Step S192), and the process proceeds to step S140. The evaluation index extraction unit 130 may derive a plurality of evaluation index groups MKS in step S130. In this case, the index value calculation unit 140 calculates the total index value MKVT by using the evaluation index group MKS with the highest separation accuracy among the plurality of evaluation index group MKS derived by the evaluation index extraction unit 130. When there are a plurality of evaluation index group MKSs having high separation accuracy, the index value calculation unit 140 uses the evaluation index group MKSs having the same separation accuracy as the evaluation index group MKS used in the previous evaluation to obtain the total index value. MKVT may be calculated.

Here, since the evaluation index MKS used for each evaluation is different from each other, even if the evaluation result by a certain evaluation index group MKS is a negative example, the evaluation result by a different evaluation index group MKS is a positive example. On the contrary, even if the evaluation result by one evaluation index group MKS is a positive example, the evaluation result by a different evaluation index group MKS may be a negative example. Therefore, by evaluating the target cell TC using a plurality of evaluation index groups MKS by the above-mentioned treatment, the cell evaluation device 10 can evaluate various target cell TCs.

In the above description, a case where re-evaluation using another evaluation index group MKS is performed based on the input to the user's input device 50 has been described, but the present invention is not limited to this. For example, when the evaluation unit 150 evaluates that a negative case is established, the cell evaluation device 10 may perform re-evaluation using another evaluation index group MKS, and the evaluation unit 150 gives a positive example. Even if it is evaluated that it holds, it may be re-evaluated using another evaluation index group MKS.

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment and may be appropriately modified without departing from the spirit of the present invention. it can. The configurations described in each of the above-described embodiments may be combined.

10 ... Cell evaluation device, 110 ... Target image acquisition unit, 120 ... Evaluation condition acquisition unit, 130 ... Evaluation index extraction unit, 140 ... Index value calculation unit, 150 ... Evaluation unit, 160 ... Display control unit

Claims (21)

1. An evaluation condition acquisition unit that acquires evaluation conditions for cells selected by the user,
   An evaluation index obtained from an image containing the cells based on the evaluation conditions acquired by the evaluation condition acquisition unit, and a value representing the accuracy of separating positive and negative examples of the evaluation conditions is predetermined. An evaluation index selection unit that selects an evaluation index having a value equal to or higher than the value and specifies a suitable time for acquiring an image used for calculating the evaluation index, and an evaluation index selection unit.
   An evaluation unit that evaluates the cells for the evaluation conditions based on the evaluation index selected by the evaluation index selection unit in the image acquired at the specified time.
   A cell evaluation device comprising.
2. An index value calculation unit for calculating an index value of the evaluation index based on the image acquired at the specified time based on the image is further provided.
   The evaluation unit evaluates the cells with respect to the evaluation conditions based on the index value calculated by the index value calculation unit.
   The cell evaluation device according to claim 1.
3. The evaluation index selection unit specifies the time based on a value representing the accuracy of separating the positive example and the negative example of the evaluation condition.
   The cell evaluation device according to claim 1 or 2.
4. The value representing the accuracy of separating the positive example and the negative example of the evaluation condition is the AUC value of the receiver operating characteristic curve derived from the index value of the positive example group and the negative example group of the evaluation condition.
   The cell evaluation device according to claim 2 or 3.
5. The value representing the accuracy of separating the positive example and the negative example of the evaluation condition is the value calculated by the z-test from the significance probability based on the index value of the positive example group and the negative example group of the evaluation condition, and the sample size. The value divided by the square root,
   The cell evaluation device according to claim 2 or 3.
6. The evaluation conditions include evaluation of the quality of cell yield, and the evaluation conditions include evaluation.
   The cell evaluation device according to any one of claims 1 to 5, wherein the positive and negative examples of the evaluation conditions are the quality of the cell yield.
7. The evaluation conditions include evaluating the quality of the speed at which the cells of interest are cultured.
   The positive and negative examples of the evaluation conditions are the quality of the speed at which the target cells are cultured.
   The cell evaluation device according to any one of claims 1 to 6.
8. The evaluation condition includes evaluating whether or not the cell is a cell to be cultured.
   Positive and negative examples of the evaluation conditions are whether or not the cells are cells to be cultured.
   The cell evaluation device according to any one of claims 1 to 7.
9. The evaluation conditions include evaluating whether or not the procedure performed on the cells is uniform.
   The positive and negative examples of the evaluation conditions are whether or not the procedure is uniform.
   The cell evaluation device according to any one of claims 1 to 8.
10. When the evaluation unit evaluates that the negative example is established as a result of evaluating the cells, the evaluation unit evaluates the cells again using an evaluation index different from the evaluation index, whichever is one of claims 1 to 9. The cell evaluation device according to one item.
11. On the computer
    Get cell evaluation conditions,
    An evaluation index obtained from an image containing the cells based on the acquired evaluation conditions, wherein the value representing the accuracy of separating the positive and negative examples of the evaluation conditions is equal to or higher than a predetermined value. Is selected, and a suitable time for acquiring an image used for calculating the evaluation index is specified.
    The cells are evaluated for the evaluation conditions based on the selected evaluation index in the image acquired at the specified time.
    program.
12. Obtaining cell evaluation conditions and
    An evaluation index obtained from an image containing the cells based on the acquired evaluation conditions, wherein the value representing the accuracy of separating the positive and negative examples of the evaluation conditions is equal to or higher than a predetermined value. And to specify a suitable time to acquire an image used for calculating the evaluation index.
    To evaluate the cells for the evaluation conditions based on the selected evaluation index in the image acquired at the specified time.
    Cell evaluation method including.
13. Based on the image, the index value of the evaluation index based on the image acquired at the specified time is calculated.
    Evaluating the cells means evaluating the cells for the evaluation conditions based on the calculated index value.
    The cell evaluation method according to claim 12.
14. The time is specified based on a value representing the accuracy of separating the positive and negative examples of the evaluation condition.
    The cell evaluation method according to claim 12 or 13.
15. The value representing the accuracy of separating the positive example and the negative example of the evaluation condition is the AUC value of the receiver operating characteristic curve derived from the index value of the positive example group and the negative example group of the evaluation condition.
    The cell evaluation method according to any one of claims 12 to 14.
16. The value representing the accuracy of separating the positive example and the negative example of the evaluation condition is the value calculated by the z-test from the significance probability based on the index value of the positive example group and the negative example group of the evaluation condition, and the sample size. The value divided by the square root,
    The cell evaluation method according to any one of claims 12 to 14.
17. The evaluation conditions include evaluation of the quality of cell yield, and the evaluation conditions include evaluation.
    The positive and negative examples of the evaluation conditions are the quality of the yield of the cells.
    The cell evaluation method according to any one of claims 12 to 16.
18. The evaluation conditions include evaluating the quality of the speed at which the cells of interest are cultured.
    The positive and negative examples of the evaluation conditions are the quality of the speed at which the target cells are cultured.
    The cell evaluation method according to any one of claims 12 to 17.
19. The evaluation condition includes evaluating whether or not the cell is a cell to be cultured.
    Positive and negative examples of the evaluation conditions are whether or not the cells are cells to be cultured.
    The cell evaluation method according to any one of claims 12 to 18.
20. The evaluation conditions include evaluating whether or not the procedure performed on the cells is uniform.
    The positive and negative examples of the evaluation conditions are whether or not the procedure is uniform.
    The cell evaluation method according to any one of claims 12 to 19.
21. In evaluating the cells, when it is evaluated that a negative example is established as a result of evaluating the cells, the cells are evaluated again using an evaluation index different from the evaluation index.
    The cell evaluation method according to any one of claims 12 to 20.

Images