Information acquisition method for acquiring information on cell viability, information acquisition system, information acquisition program, and recording medium

Abstract

To enable acquisition of information on cell viability based on tracking information for each cell in cell images taken over time and the results of determining the viable cell area.SOLUTION: An information acquisition method comprises: an image acquisition step for acquiring multiple cell images captured at multiple different imaging times; a live cell region determination step for determining whether a cell region present in the multiple cell images is a live cell region containing a live cell; a tracking step for performing a correspondence between live cell regions originating from the same cell region present in each of the multiple cell images captured at the multiple different imaging times, starting from a live cell region present in a cell image captured at a second imaging time, tracing the correspondence back in time to the first imaging time; and a viability determination step for determining, from among the cell regions present in the cell image captured at the first imaging time, live cell regions tracked in the tracking step as viable.SELECTED DRAWING: None

Description

[Technical Field]

[0001] The disclosure of the present specification relates to an information acquisition method, an information acquisition system, an information acquisition program, and a recording medium for acquiring information related to cell viability. [Background technology]

[0002] In the fields of cell culture and cell manufacturing, determining the viability of cells and counting the number of viable cells are extremely important and are routinely performed in the culture, manufacturing, and quality control processes. Currently, the most common method involves taking images of a cell sample for evaluation that has been stained with a dead cell staining reagent such as trypan blue, determining that unstained cells are live cells and stained cells are dead cells, and counting the number of each.

[0003] In light of the fact that the state of cells changes from moment to moment and the number of live and dead cells changes over time, Patent Document 1 discloses a technique for capturing time-lapse images of cells and determining whether each cell region is live or dead. Furthermore, it discloses a technique for matching cell regions in each image with each other and determining cells as dead when the proportion of cells determined as dead is equal to or greater than a certain level. Furthermore, Patent Document 2 discloses a technique in which time-lapse images of cells are taken at two different times, and the frequency of cell death events at each time is analyzed and compared. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2006-314214 [Patent Document 2] Japanese Patent Application Publication No. 2017-023055 Summary of the Invention [Problem to be solved by the invention]

[0005] With conventional techniques, it has been difficult to obtain information regarding cell viability. In view of the above problems, the present invention aims to provide an information acquisition method, an information acquisition system, an information acquisition program, and a recording medium for acquiring information on the viability of cells contained in a cell stock or the like. [Means for solving the problem]

[0006] According to the present invention, an image acquisition step of acquiring a plurality of cell images captured at a plurality of different imaging timings including a first timing and a second timing; a live cell region determination step of determining whether a cell region present in each of the plurality of cell images is a live cell region containing a live cell; a tracking step of tracking live cell regions originating from the same cell region by tracing back in time to the first timing, starting from a live cell region present in the cell image captured at the second timing, and associating live cell regions originating from the same cell region present in each of the plurality of cell images captured at the plurality of different imaging timings; An information acquisition method for acquiring information regarding cell viability is provided, characterized by having a viability determination step of determining that a live cell region that was tracked in the tracking step is viable among the cell regions present in the cell image captured at the first timing.

[0007] According to the present invention, there is also provided an information acquisition program for causing a computer to execute the information acquisition method, and a recording medium storing the information acquisition program in a computer-readable format.

[0008] Furthermore, according to the present invention, an image acquisition unit that acquires a plurality of cell images captured at a plurality of different imaging timings including a first timing and a second timing; a live cell region determination unit that determines whether a cell region present in each of the plurality of cell images is a live cell region containing a live cell; a tracking unit that tracks live cell regions originating from the same cell region by tracing back in time to the first timing, starting from a live cell region present in the cell image captured at the second timing, and associating live cell regions originating from the same cell region present in each of the plurality of cell images captured at the plurality of different imaging timings; An information acquisition system for acquiring information regarding cell viability is provided, characterized by having a viability determination unit that determines, among the cell regions present in the cell image captured at the first timing, live cell regions that can be tracked by the tracking unit as viable. [Effects of the Invention]

[0009] According to the disclosure of this specification, information on cell viability can be obtained based on the tracking information of each cell in cell images captured over time and the results of live cell region determination, which can provide more useful information to cell users, particularly in quality inspection of large lots of cells such as cell stocks. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a flow chart showing an example of an information acquisition method according to the present invention. [Figure 2] 1 is a diagram showing an example of the configuration of an information acquisition system according to the present invention. [Figure 3] FIG. 1 is a diagram showing an example of the hardware configuration of an information acquisition system according to the present invention. [Figure 4(a)] FIG. 10 is a diagram showing an example of a processing method for a live cell region determination step. [Figure 4(b)] FIG. 10 is a diagram showing an example of data output in the live cell region determination step. [Figure 5] 10 is a flow diagram showing an example of a cell region extraction process. [Figure 6] 10 is a flow chart showing a first example of a processing method of a tracking step. [Figure 7(a)] FIG. 10 is a diagram showing a first example of a processing method in a tracking step. [Figure 7(b)]FIG. 10 is a diagram showing a first example of data output in a tracking process. [Figure 8(a)] FIG. 10 is a diagram showing a second example of a processing method in a tracking step. [Figure 8(b)] FIG. 10 is a diagram showing a second example of data output in the tracking process. [Figure 9(a)] FIG. 10 is a diagram showing an example of a process for determining a live cell region in a cell image captured after staining live or dead cells and adding a reagent that absorbs visible light. [Figure 9(b)] FIG. 10 is a diagram showing an example of data output in the live cell region determination step. [Figure 10] 10A and 10B are diagrams showing an example of a process for determining a live cell region in a cell image captured after staining live or dead cells and adding a fluorescent reagent. [Figure 11(a)] FIG. 10 is a diagram showing an example of a live cell region determination process using a pre-trained classifier. [Figure 11(b)] FIG. 10 is a diagram showing an example of data output in the live cell region determination step. [Figure 12] (a) to (f) are diagrams showing a first example of a processing method for the tracking step when fusion or separation of cell regions occurs. [Figure 13] (a) to (f) are diagrams showing a second example of a processing method for the tracking step when fusion or separation of cell regions occurs. [Figure 14] FIG. 10 shows a first example of a UI screen that displays the results of viability assessment together with image data. [Figure 15] FIG. 10 shows a second example of a UI screen displaying the viability assessment results together with image data. [Figure 16] FIG. 10 shows a third example of a UI screen displaying the viability assessment results together with image data. [Figure 17] FIG. 10 is a diagram showing the results of a comparison between the correlation between survival rate and confluency and the correlation between viable cell rate and confluency in Examples and Reference Examples. [Figure 18] FIG. 1 is a graph showing the correlation between survival number and confluency in Examples and Reference Examples. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the invention according to the claims. Although multiple features are described in the embodiments, not all of these multiple features are necessarily essential to the invention, and multiple features may be combined arbitrarily. Note that in the drawings, similar components are given the same reference numerals, and duplicated explanations will be omitted.

[0012] (Problems to be solved by the information acquisition method according to this embodiment) The number of live cells is counted as one item in the quality inspection of cell stocks, etc., but cell users may want to know not the number of live cells contained in the cell stock, but the number of cells that have the ability to continue living, i.e., the number of viable cells. However, counting live cells using dead cell stains such as trypan blue has the problem of misidentifying cells that die later, i.e., non-viable cells, as live cells. Furthermore, Patent Document 1 improves detection accuracy by making a comprehensive judgment based on multiple live / dead cell determination results obtained over time, thereby suppressing erroneous detection of dead cells due to temporary circularization of cell region images, etc., but it has been difficult to obtain information about cell viability.

[0013] FIG. 1 is a flow chart illustrating an example of an information acquisition method for acquiring information about cell viability according to this embodiment. The information acquisition method according to this embodiment includes step S101, which is an image acquisition step for acquiring multiple cell images captured at multiple different imaging times, including a first imaging time and a second imaging time. It also includes step S102, which is a live cell region determination step for determining whether a cell region present in each of the multiple cell images contains a live cell. It also includes step S103, which is a tracking step for tracking live cell regions originating from the same cell region, by starting from a live cell region present in the cell image captured at the second imaging time and chronologically tracing back to the first imaging time, to associate live cell regions originating from the same cell region present in each of the multiple cell images captured at different imaging times. Finally, it includes step S104, which is a viability determination step for determining live cell regions tracked in the tracking step among the cell regions present in the cell image captured at the first imaging time as viable. The method may further include step S105, which is a viability quantification calculation process that calculates at least one of the number and the proportion of live cell areas that are determined to be viable in the viability determination process among the cell areas present in the cell image captured at the first timing.

[0014] FIG. 2 shows an example of the configuration of an information acquisition system 201 and other devices according to this embodiment. The information acquisition system 201 according to this embodiment includes an image acquisition unit 202 that acquires multiple cell images captured at multiple different capture times, including a first capture time and a second capture time. It also includes a live cell region determination unit 203 that determines whether a cell region present in each of the multiple cell images is a live cell region containing a live cell. It also includes a tracking unit 204 that tracks live cell regions originating from the same cell region by starting from a live cell region present in a cell image captured at the second capture time and chronologically tracing back to the first capture time to associate live cell regions originating from the same cell region present in each of the multiple cell images captured at different capture times. It also includes a viability determination unit 205 that determines live cell regions tracked by the tracking unit 204 as viable among the cell regions present in the cell images captured at the first capture time. The information acquisition system 201 may also include a memory unit 206 and a viability quantification calculation unit 207, which will be described later. The image acquisition unit 202, viable cell region determination unit 203, tracking unit 204, viability determination unit 205, and viability quantification calculation unit 207 can each perform steps S101 to S105 described in the following embodiments and modifications. The storage unit 206 stores intermediate and final output information for each step, control programs and analysis programs for executing the processing of each step, and predetermined rules and trained classifiers for determining whether or not a region is a viable cell. The storage unit 206 does not necessarily have to be provided within the system, and may be provided externally as a storage device, or may be provided externally as an input / output device capable of reading and writing data from and to a recording medium.

[0015] The information acquisition system 201 is externally connected to a data server 208, an input device 211, and an output device 212. Furthermore, an observation device 210 is connected to the data server 208.

[0016] The observation device 210 is a device that captures cell images such as phase contrast images, dark field images, differential interference contrast images, transmitted light images, and fluorescent images. To capture images over time during culture, an automated cell culture observation system may be used in which a cell culture incubator and the observation device 210 are integrated and the imaging schedule is programmable. The observation device 210 is connected to a data server 208 and outputs image data to the data server 208 to store the captured cell images.

[0017] The data server 208 stores time-lapse cell image data (a plurality of cell images captured at a plurality of different capture timings including a first timing and a second timing) captured by the observation device 210. The data server 208 is, for example, a recording medium such as a hard disk drive (HDD) or a solid state drive (SSD) connected to the observation device 210 and the information acquisition system 201 via a network.

[0018] The input device 211 is a device for inputting information to the information acquisition system 201, and is typically a user interface for a user to operate the information acquisition system 201. Examples of the input device 211 include a keyboard, a button, a mouse, a touch panel, and the like.

[0019] The output device 212 is a device that outputs information from the information acquisition system 201 to the outside, and is typically a user interface for presenting information to a user. Examples of the output device 212 include a display, a printer, and a speaker. The input device 211 and the output device 212 may be integrally formed as a touch panel.

[0020] The configuration of the information acquisition system 201 according to this embodiment is not limited to the above configuration, and for example, the information acquisition system 201 may include some or all of the observation device 210, the data server 208, the input device 211, and the output device 212 therein.

[0021] 3 is a block diagram showing an example of the hardware configuration of an information acquisition system 201 according to this embodiment. The information acquisition system 201 has a CPU 301, a RAM 302, a ROM 303, a HDD 304, and a communication I / F 305. These components are connected to each other via a bus or the like.

[0022] The CPU 301 is a processor that reads out programs stored in the ROM 303 and the HDD 304 into the RAM 302 and executes them to perform calculations and control each part of the information acquisition system 201. The processing performed by the CPU 301 may include extraction of a cell region, determination of whether or not it is a live cell region, determination of viability, calculation of quantitative information regarding viability, and the like.

[0023] The RAM 302 is a volatile storage medium and functions as a work memory when the CPU 301 executes programs. The ROM 303 is a non-volatile storage medium and stores firmware and the like necessary for the operation of the information acquisition system 201. The HDD 304 is a non-volatile storage medium and stores programs used in processes such as extracting a cell region, determining whether or not a region is a live cell region, determining viability, and calculating quantitative information related to viability in this embodiment, as well as image data and the like.

[0024] The communication I / F 305 is a communication device based on standards such as Wi-Fi (registered trademark), Ethernet (registered trademark), Bluetooth (registered trademark), etc. The communication I / F 305 is used for communication with the observation device 210, other computers, etc.

[0025] The hardware configuration of the information acquisition system 201 described above is an example and can be modified as appropriate. Devices other than those described above may be added, or some devices may not be provided. Also, some devices may be replaced with other devices having similar functions. Furthermore, some functions may be provided by other devices via a network, and the functions constituting this embodiment may be distributed and realized among multiple devices.

[0026] Furthermore, examples of processors that can be installed in the information acquisition system 201 include, in addition to the above-mentioned CPU 301, a GPU, an ASIC, an FPGA, etc. Furthermore, a plurality of these processors may be provided, and processing may be performed by the plurality of processors in a distributed manner. Furthermore, the function of storing information such as image data in the HDD 304 may be provided in a data server rather than within the information acquisition system 201. Furthermore, the HDD 304 may be a storage medium such as an optical disk, a magneto-optical disk, or an SSD.

[0027] First Embodiment (overview) In this embodiment, a method for acquiring information on cell viability based on tracking information of each cell region in cell images captured over time and the results of determining the viable cell region using the information acquisition method and information acquisition system of this embodiment will be described.

[0028] (Step S101: Image acquisition process) In this embodiment, step S101 is an image acquiring step in which the image acquiring unit 202 acquires a plurality of cell images captured at a plurality of different imaging timings, including a first timing and a second timing that is later than the first timing. The imaging timings may further include one or more timings between the first timing and the second timing. That is, this is an image acquisition step for acquiring a plurality of cell images captured over time. In this step, the image acquisition unit 202 can acquire the cell images captured by the observation device 210 and stored in the data server 208 in response to a processing request from a user. The image acquisition unit 202 is connected to a data server 208 via a network or the like, and can acquire multiple cell images stored in the data server 208.

[0029] The cell image acquired in this step is an image of a cell from which information about viability (whether or not the cell is viable) is to be acquired, and it is sufficient if it can identify a cell region containing cells and determine whether or not the cell region contains live cells. The state of the target cells may be either a single-cell state in which individual cells are separated, or an aggregated state such as a colony in which multiple cells are gathered. Which state prevails depends on the type of cell and the culture method. Note that the cell region here may refer to either a region consisting of a single cell or a region consisting of multiple cells, and includes both a cell region determined to be a live cell region containing live cells in the live cell region determination step described below, and a cell region not determined to be a live cell region (dead cell region).

[0030] Cell images suitable for unstained cells include phase contrast images, dark field images, and differential interference contrast images. Cell images for stained cells include transmitted light images and fluorescent images. For time-lapse imaging during culture, an automated cell culture observation system that integrates a cell culture incubator and observation device and allows for a programmable imaging schedule is ideal, but imaging and cell sample movement can also be performed manually.

[0031] The cell images are acquired by capturing images over time at different capture times, from a first capture time at which a cell image including a cell region to be assessed for viability is captured to a subsequent second capture time. For example, when assessing viability as an examination of a cell stock, it is preferable to set the first capture time as the time when the cell stock is opened and culture is initiated, and the second capture time as the time when culture has stabilized and dead cells are less likely to occur.

[0032] Since it is necessary to match the cell regions (living cell regions) between cell images captured over time in a later process, it is desirable to measure the migration and movement speed of the cells in advance and set the image capture interval short enough to enable matching. The number of images to be captured can be determined from the interval between the first and second timings, and the interval between the images captured between the first and second timings.

[0033] Images captured over time may be acquired in real time from the observation device 210 and / or the data server 208, for example, each time a photograph is taken, or may be stored in an external storage area such as a HDD or cloud storage and then acquired all at once later.

[0034] Each cell image may be stored in advance in the memory unit 206 of the information acquisition system 201, and the image acquisition unit 202 may acquire each cell image stored in advance from the memory unit 206.

[0035] In this step, the image acquiring unit 202 acquires cell image data captured at a plurality of different capture timings from the first timing to the second timing. At this time, it is preferable that the image acquiring unit 202 assigns a timing ID indicating the capture timing to each cell image. For example, the image acquiring unit 202 acquires a display image as a cell image and image information related to the display image from the observation device 210 and / or the data server 208, and assigns a timing ID to the cell image based on information such as the capture time of the cell image included in the image information. More specifically, the timing IDs t1 to t arranged in order of the capture time or elapsed time of the cell images are assigned to the cell images. N (N is the number of times of shooting, N is 2 or more) is used as the timing ID. In this case, the timing ID of the first timing is t1, and the timing ID of the second timing is t N This becomes:

[0036] Each cell image and timing ID acquired by the image acquisition unit 202 may be transmitted to the live cell region determination unit 203 which performs the next step, or may be stored in the memory unit 206 as appropriate and acquired by the live cell region determination unit 203.

[0037] (Step S102: Viable cell region determination step) In this embodiment, step S102 is a live cell region determination process in which the live cell region determination unit 203 extracts cell regions from each cell image acquired by the image acquisition unit 202 in step S101 and determines whether each cell region is a live cell region containing a live cell. Figure 4(a) shows the process diagrammatically.

[0038] First, images are taken at different timings, and each timing ID (t1 to t N ) are assigned to a plurality of cell images 401, and a cell region 402 is extracted from the cell images 401. A known image analysis technique using image processing or machine learning can be used to extract the cell region 402. In this embodiment, the cell region 402 is extracted using a differential filter and binarization processing. Hereinafter, the cell region extraction step for extracting the cell region 402 will be described with reference to FIG. 5.

[0039] Step S501 is a differential image generation step in which a differential image is generated by applying a differential filter to a cell image. A differential image is an image in which the amount of change in brightness between each pixel and its surrounding pixels is calculated and expressed as an image. In the case of a cell image, the differential image has high brightness values ​​at the outline of the cell region and at the outline of the cells within the cell region.

[0040] Step S502 is a binarization process step in which a binarization process is performed on the differential image generated in step S501, thereby extracting areas having high luminance values ​​in the differential image. In the binarization process, an arbitrary threshold is set, and the value of each pixel in the differential image is replaced with a value of 1 if it is equal to or greater than the threshold, and with a value of 0 if it is less than the threshold.

[0041] Here, the method of binarization is not limited to a method of setting an arbitrary threshold. For example, a method of automatically determining a threshold, such as Otsu's binarization or Li's binarization, may be used. When an arbitrary threshold is set, the threshold is set according to the imaging conditions of the device, such as exposure time and focus setting. Alternatively, a method of determining a threshold for each pixel of an image, such as adaptive binarization, may be used. The binarization process creates a binary image (hereinafter referred to as an edge image) in which pixel values ​​in areas with large changes in brightness are expressed as 1 and other pixel values ​​as 0.

[0042] Step S503 is a mask image generation process that generates a mask image of the cellular region based on the edge image generated in step S502. Here, the mask image is a binary image in which the cellular region is represented by a pixel value of 1 and other regions by a pixel value of 0. The mask image is generated by extracting regions in the edge image where pixel values ​​of 1 are connected, and replacing the pixel values ​​inside each connected region with 1.

[0043] Step S504 is a labeling process step in which region coordinate information is acquired from the mask image generated in step S503, and a labeling process is performed to generate a labeled image to which a region ID for identifying each cell region is assigned. In the labeling process step, regions in the mask image where pixel values ​​of 1 are connected are extracted, and for each connected region, the pixels in that region are replaced with different pixel values.

[0044] By performing the processes from step S501 to step S504 on the cell image of each timing ID acquired in step S101, the cell region 402 in each cell image 401 is extracted. Also, as shown in Fig. 4(a), the live cell region determination unit 203 assigns a region ID (C1 to C5) to each cell region 402 present in each cell image 401.

[0045] Here, for example, when the cells to be determined as being a viable cell region form an aggregate of multiple cells, such as a colony, each cell may be extracted as a separate cell region 402, or the entire aggregate may be extracted as one cell region 402. Either extraction method can be selected according to the purpose by optimizing the technique, threshold setting, etc. in the extraction process described above.

[0046] Next, it is determined whether the extracted cell region 402 is a live cell region containing live cells or not. For example, this determination is made according to a pre-set rule based on the image features of each cell region. When a cell region 402 consists of multiple cells, it is not necessary for all cells to be live cells; if a live cell is included, it is determined to be a live cell region. The determination of whether or not it is a live cell region is performed on all cell regions extracted from the cell image of each timing ID. As a result, as shown in Figure 4(a), it is divided into a live cell region 403 and a cell region 404 that was not determined to be a live cell region.

[0047] Finally, as shown in Figure 4(b), for each cell region contained in each cell image, a timing ID that identifies the timing of the image capture, a region ID that identifies the cell region, region coordinate information, and information indicating whether the region is a live cell region are output, for example, in table format.

[0048] The above timing ID, region ID, region coordinate information, and information indicating whether it is a live cell region may be transmitted from the live cell region determination unit 203 to the tracking unit 204 which performs the next step, or may be stored in the memory unit 206 as appropriate and acquired by the tracking unit 204.

[0049] (Step S103: Tracking process) In this embodiment, step S103 is a tracking process in which the tracking unit 204 tracks live cell regions originating from the same cell region by starting from a live cell region present in a cell image captured at a second timing and sequentially associating live cell regions originating from the same cell region that exist in each of multiple cell images captured at multiple different imaging timings, going back in time to the first timing.

[0050] Figures 6 and 7(a) show an example of this flow. First, in step S601 shown in FIG. N (second timing) and t N-1 Of the cell region 402 included in the cell image 401 captured at the capture timing (one before the second timing), the region ID and region coordinate information are acquired for the live cell region 403. where t N A new unique live cell tracking ID (L1, L2) is assigned to each live cell region 403 in the cell image 401. N Alternatively, the same region ID as that of each living cell region 403 in the cell image 401 may be used.

[0051] And t N and t N-1 For the live cell regions 403 present in the cell images 401 captured at the timing of capturing, correspondence is established between live cell regions originating from the same cell region. As a method of correspondence, a method of determining whether the live cell regions originate from the same cell region based on the amount of overlap between the live cell regions of the two cell images can be used. For example, N The region ID of the live cell region in the cell image is A1, t N-1The area ID of the live cell area is represented as A2, the areas of A1 and A2 are represented as S(A1) and S(A2), respectively, and the area of ​​overlap between the two is represented as S(A1∩A2). The amount of overlap is calculated using the following formula (1), and live cell areas with the largest overlap amount greater than 0 can be determined to originate from the same cell area and associated. The lower limit of the amount of overlap can be set to any value depending on the type of cell, etc.

number

[0052] In this embodiment, first, in step S601, the second timing t N Starting from the live cell region in the cell image, N-1 By tracing back the cell images, we can match live cell regions originating from the same cell region. N For the region ID of a specific living cell region to be matched among the cell images, t N-1 The region IDs of the live cell regions in the cell image are sequentially changed, and N t determined to originate from the same cell region as a specific live cell region of N-1 If it is determined that the living cell regions are derived from the same cell region, the tracking unit 204 searches for the living cell region of t N-1 Live cell regions and t in cell images N The same live cell tracking ID is assigned to the live cell area in the cell image. N and t N-1 The correspondence between specific living cell regions in each cell image is completed. N By repeating this process for all live cell regions in the cell image, t N For the whole live cell area in the cell image, t N-1 The living cell regions can be traced back.

[0053] The above determination method is merely an example. Other methods, such as determining the distance between centers of gravity or determining optical flow, such as the Lucas-Kanade method, may also be used to distinguish live cell regions originating from the same cell region. Alternatively, tracking may be performed by generating 3D volume data by chronologically overlaying mask images of the cell regions created at each imaging timing, and then detecting connected regions in the 3D volume data. Detection of connected regions can be achieved by applying the labeling process performed in step S504 to the 3D volume data. The labeling process for the 3D volume data assigns the same value to regions connected in the width, height, and time directions of the captured cell images. By assigning this value as a live cell tracking ID for the cell region at each imaging timing, correspondence between live cell regions can be established.

[0054] Steps S602 to S604 show a flow in which this association work is repeated going back in time series in order up to the first timing. N If N is 2, the tracking process is terminated. If N is greater than 2, the following step S603 is performed. Step S603 is a step in which the timing ID is t k (k is 2 to N-1) k+1 The live cell region that has been associated with the live cell region of the cell image of t and the timing ID of t k-1 This is performed in step S604 by associating the cell image with the living cell region. k-1 = t1. That is, the process of associating a live cell region included in a cell image captured at the immediately previous capture timing with a live cell region in a cell image captured at a later capture timing in the most recent process is repeated. This allows live cell regions 403 originating from the same cell region to be associated and tracked in chronological order going back to the first timing, as shown in FIG. 7(a). The associated live cell regions 403 are assigned the same live cell tracking ID (L1, L2).

[0055] Finally, as a result of the above-mentioned association and tracking, the timing ID, region ID, region coordinate information, information indicating whether or not the region is a live cell region, and live cell tracking ID for each cell region are output as live cell region association information, as shown in Fig. 7(b). The live cell region association information including the timing ID, region ID, region coordinate information, information indicating whether or not the region is a live cell region, and live cell tracking ID may be transmitted from the tracking unit to the viability assessment unit, or may be appropriately stored in a storage unit and acquired by the viability assessment unit.

[0056] In the above, the method of associating only the live cell regions from the second timing back to the first timing has been described. As another method of associating, the method shown in FIG. 8(a) is also suitable. First, for all cell regions 402 contained in each cell image 401, correspondence is performed between cell regions originating from the same cell region between cell images taken at different times. That is, correspondence is performed between cell regions originating from the same cell region that exist in multiple cell images taken at different times. Then, cell tracking IDs (T1 to T5) are assigned so that cell regions originating from the same cell region have the same ID. Correspondence between cell regions originating from the same cell region that exist in multiple cell images can be performed by replacing "living cell region" with "cell region" in the calculation of the overlap amount using the above-mentioned formula (1). Alternatively, if tracking is performed by generating 3D volume data by overlapping mask images of cell regions created at each imaging time in chronological order and detecting connected regions, this can be performed for all cell regions and cell tracking IDs can be assigned.

[0057] Furthermore, the second timing was N The same live cell tracking IDs (L1 to L2) are assigned to live cell regions in different cell images that have the same cell tracking ID as a live cell region in the cell image in chronological order, thereby associating the live cell regions with each other.

[0058] In detail, the correspondence between live cell regions is as follows: first, the timing ID is t N and t N―1 In the cell image 401, the same live cell tracking ID is assigned to live cell regions 403 that originate from the same cell region and have the same cell tracking ID. This allows for correspondence between live cell regions that originate from the same cell region (FIG. 6, step S601). N If N is 2, the tracking process is terminated. If N is greater than 2, step S603 is performed. In step S603, as described above, if the timing ID is t k The cell image of t k+1 The live cell region of the cell image of t and the live cell region that has been associated with the live cell region of t k-1 The living cell region in the cell image is then matched with the living cell region in the cell image. k-1 Repeat until t=t1 (S604).

[0059] That is, after matching cell regions originating from the same cell region that exist in a plurality of cell images with each other, starting from a live cell region that exists in a cell image captured at a second timing, matching cell regions originating from the same cell region that exist in a plurality of cell images is performed by tracing back in time series to the first timing. Note that in this embodiment, a case is shown in which matching cell regions originating from the same cell region is performed for all cell regions in chronological order from t1, which is the first timing, in the order of t2 → t3 → ..., but the order of matching cell regions is not limited to this.

[0060] In this case, as shown in Fig. 8(b), the timing ID, region ID, region coordinate information, information indicating whether the region is a live cell region, cell tracking ID, and live cell tracking ID for each cell region are output as correspondence information for the live cell region. The timing ID, region ID, region coordinate information, information indicating whether the region is a live cell region, cell tracking ID, and live cell tracking ID may be transmitted from the tracking unit to the viability assessment unit, or may be stored in a storage unit as appropriate and acquired by the viability assessment unit.

[0061] (Step S104: Viability Determination Step) Step S104 is a viability determination step in which the viability determination unit 205 determines that a live cell region that was tracked in the tracking step among the cell regions present in the cell image captured at the first timing is viable.

[0062] The process in this step is as follows: In step S103, it is determined whether a live cell tracking ID has been assigned to a cell region present in the cell image of timing ID (t1) corresponding to the first timing, based on the live cell region association information output by tracking unit 204. A live cell region to which a live cell tracking ID has been assigned is determined to be "viable," and a cell region to which no live cell tracking ID has been assigned (a live cell region or a cell region not determined to be a live cell region) is determined to be "non-viable."

[0063] The information on cell viability thus obtained is added to the corresponding information on live cell regions output by tracking unit 204 in step S103 and output. Viability determination unit 205 outputs the corresponding information on live cell regions and the information on cell viability to output device 212 via a network or the like, and can present the determination results to the user. Furthermore, the information on cell viability (and the corresponding information on live cell regions) may be transmitted from viability determination unit 205 to a unit that performs the next step (for example, viability quantification calculation unit 207 in Modification 1), or may be stored in storage unit 206 as appropriate and acquired by a unit that performs the next step.

[0064] (Modification 1 (Step S105): Viability Quantity Calculation Step) In quality inspections of cell stocks and the like, it is important to acquire quantitative information regarding cell viability. Therefore, the information acquisition system 201 preferably includes a viability quantification calculation unit 207, and the information acquisition method preferably includes a viability quantification calculation step S105. In this step, the viability quantification calculation unit 207 acquires information regarding cell viability obtained in step S104 (and information associating live cell regions obtained in step S103). Then, among the number of cell regions present in the cell image captured at a first timing, the unit 207 calculates at least one of the number and the proportion of live cell regions determined to be viable, thereby calculating quantitative information regarding cell viability. The proportion may be the proportion of the number of live cell regions determined to be viable among the number of cell regions present in the cell image captured at the first timing, or the proportion of the number of live cell regions determined to be viable among the number of live cell regions present in the cell image captured at the first timing.

[0065] In this modification, the viability quantification calculation step is performed by the viability quantification calculation unit 207, but it may also be performed by the viability determination unit 205 or by a unit that performs another step.

[0066] The viability quantification calculation unit 207 can output the correspondence information of the viable cell regions, information regarding the cell viability, and the number and / or proportion of the viable cell regions determined to be viable to the output device 212 via a network or the like, and present the results to the user.

[0067] (Variation 2: Determining the viable cell region by staining) In step S101, which is the image acquisition step, a cell image is acquired by photographing cells that have been stained with a reagent that stains at least one of live cells and dead cells, and in step S102, which is the live cell region determination step, the viable cell region is determined based on the degree of staining. Examples of reagents that preferentially stain live cells include calcein AM, fluorescent esterase reagents including fluorescein diacetate derivatives, and other thiol-reactive fluorescent reagents. Examples of reagents that preferentially stain dead cells include trypan blue, propidium iodide, 7-amino-actinomycin D, and ethidium homodimer. It is also preferable to use a combination of multiple reagents that stain at least one of live cells and dead cells, such as Hoechst-based nuclear staining reagents such as Hoechst 33342 and various organelle staining reagents. These can be easily selected by those skilled in the art.

[0068] For example, as shown in FIG. 9(a), before the first timing, live or dead cells are stained, a reagent that absorbs visible light is added, and the cells are photographed using transmitted light to obtain a cell image 401 in step S101. In this case, a stained cell region (901) and an insufficiently stained cell region 902 appear on the cell image 401. In the live cell region determination step of step S102, the degree of staining is calculated for each extracted cell region 402. For example, the degree of staining may be calculated by calculating the average brightness or minimum brightness of the cell region in a monochrome cell image, or color information in a color cell image. When a cell region includes multiple cells, the proportion of the stained portion of the cell region may be calculated. The degree of staining is compared with a predetermined threshold, and cell regions with a degree of staining equal to or greater than the threshold are designated as stained cell regions 901, and cell regions with a degree of staining less than the threshold are designated as insufficiently stained cell regions 902, thereby determining whether each cell region is a live cell region. Then, information including the degree of staining and information indicating whether or not the region is a live cell region, as shown in FIG. 9(b), is output.

[0069] 10, for example, before the first timing, live or dead cells are stained, a fluorescent reagent is added, and a cell image 401 is acquired in step S101. In this case, in addition to a cell morphology image (1001) capturing the morphology of the cell, such as a phase contrast image, dark field image, differential interference image, or transmitted light image, a fluorescent cell image 1002 is also acquired as the cell image 401. The stained cell region 1003 can be confirmed from the fluorescent cell image 1002. Since cell images capturing morphology are often suitable for extracting cell regions 402, each cell region 402 is extracted from the cell morphology image 1001 capturing morphology, and after obtaining region coordinate information, the staining degree of the corresponding (identical) cell region in the fluorescent cell image 1002 is calculated. If a cell region includes multiple cells, the proportion of the portion of the cell region where the staining degree is equal to or greater than a certain level may be calculated together. Furthermore, the staining degree is compared with a pre-set threshold value or the like to determine whether each cell region is a live cell region, and information indicating whether it is a live cell region or not is output, as shown in Figure 9(b). Note that since there may be a misalignment between the cell morphology image 1001 capturing the cell morphology and the fluorescent cell image 1002, alignment is performed in advance as necessary.

[0070] In the explanation of this modified example, an example was given in which one type of staining reagent was used for each. Techniques such as combining two types of reagents that stain live cells and dead cells with different wavelengths, or combining reagents that stain cells, are also effective. This technique is preferable for improving the accuracy of determining whether or not a cell region is live. Multiple cell images may be taken at a single imaging timing while changing the filter or wavelength band of the excitation light source. Those skilled in the art can easily select one of these techniques.

[0071] (Variation 3: Determining Live Cell Regions Using a Pre-Trained Classifier) In order to obtain information about cell viability using cell images captured over time, it is preferable that cells are not damaged by the toxicity of staining reagents or the like during the capture of cell images. In this modification, the live cell region determination unit 203 has a classifier, and in step S102, which is the live cell region determination process, it determines whether each cell region is a live cell region as follows: The classifier is trained to input at least one of at least a portion of the cell image including the target cell region (to be determined as a live cell region) and image features calculated from at least a portion of the cell image including the target cell region, and to output the result of determining whether the target cell region is a live cell region. This makes it possible to obtain cell images captured without staining the cells in the image acquisition process, and to obtain information about cell viability without damage from the toxicity of staining reagents or the like.

[0072] The classifier uses labeled data, which is obtained using, for example, a reagent for staining live and dead cells and which is assigned information on whether a region is a live cell or not, as training data. The classifier is generated by training a machine learning model that receives as input at least one of a portion of a cell image that includes a target cell region and an image feature calculated from at least a portion of a cell image that includes the target cell region, and outputs a true / false value indicating whether a region is a live cell or not.

[0073] When the input is at least a portion of a cell image that includes the target cell region, a deep learning model such as a DNN (deep neural network), a CNN (convolutional neural network), or a GAN (generative adversarial network) is suitable as the machine learning model.

[0074] When image features calculated from at least a portion of a cell image including the target cell region are used as input, the following can be used, for example: diameter, average brightness, minimum brightness, brightness deviation, and also features representing the contour shape of the cell region (e.g., circularity) and features representing texture (e.g., homogeneity by GLCM).

[0075] The machine learning model uses a technique such as XGBoost, but is not limited to this, and models such as neural networks, SVMs (Support Vector Machines), random forests, LightGBMs, and Gaussian mixture models can also be applied.

[0076] An example of using a classifier based on image features of a cell region is shown in Figure 11(a). In the live cell region determination process of step S102, image features to be used as input to the trained classifier are calculated for each cell region 402. Using these image features as input, the classifier determines whether or not the cell region is a live cell region, and outputs information including the image features and information indicating whether or not the cell region is a live cell region, as shown in Figure 11(b).

[0077] In this modified example, the information acquisition method and information acquisition system are for acquiring information regarding cell viability without staining the cells, but it is also possible to combine modified examples 2 and 3, perform staining of the cells, and then use a classifier to determine whether or not the area is a live cell area.

[0078] (Variation 4: Tracking step when fusion or separation of cell regions occurs) When multiple cells gather and show a colony-like aggregate state, cell regions may fuse or separate. Furthermore, when individual cells show a separate single-cell state, they may separate into multiple cell regions due to cell division. In this modification, in the tracking process of step S103, live cell regions that have fused or separated between the first and second timings are associated as being of the same lineage (originating from the same cell region), thereby enabling more accurate acquisition of information regarding cell viability. Determining whether cells originate from the same cell region can be done, for example, by determining that the overlap amount calculated by Equation (1) is greater than 0, i.e., when there is overlap between cell regions (live cell regions) in the previous and next cell images, they are determined to originate from the same cell region. The lower limit of the overlap amount may be set to any value depending on the cell type, etc.

[0079] An example of a method for associating living cell regions with each other in the tracking step in this modified example will be described with reference to FIGS. 12(a) to 12(f) are cell images captured at the second timing, with timing ID t N Starting from the live cell region in the cell image, N-1 →t N-2 →... and then associate only the live cell region. k For the live cell area present in the cell image of , the timing ID of the image taken at the later timing is t k+1 The timing ID is t k The live cell region and timing ID present in the cell image are t k-1 10 shows a step of associating the image with a live cell region present in the cell image.

[0080] (a) to (c) in Fig. 12 show the timing ID of k-1 The cell image and t k This shows a case where the separation of the live cell region occurred between the timings when the two cell images with timing ID t k In the cell image, the live cell regions (hereinafter sometimes referred to as (timing ID, region ID)) with region IDs C2, C3, and C5, respectively, have already been assigned live cell tracking IDs L1, L2, and L3.

[0081] Here, the timing ID is t k Focusing on the live cell regions with region IDs C2 and C3 in the cell image, and timing ID t k-1 Figure 12(b) shows the cell image with timing ID t k The live cell regions (t k , C2), (t k , C3) and timing ID is t k-1 The live cell region (t k-1 , C4) are superimposed.k , C2) and (t k , C3) in any live cell region (t k-1 , C4) and the live cell region (t k , C2) and the live cell region (t k-1 , C4) overlapping the live cell region is greater than 0. k , C3) and the live cell region (t k-1 , C4) is also greater than 0. In such a case, as shown in FIG. 12(c), when (timing ID, region ID) is (t k , C2)(t k , C3)(t k-1 , C4) are considered to be of the same lineage (originating from the same cell region) and are assigned the same live cell tracking ID L1. k , C2)(t k , C3) were originally assigned different live cell tracking IDs (L1, L2, respectively), so they are updated to L1. k+1 ~t N For live cell regions to which a live cell tracking ID of L2 has been assigned in the timing ID up to (t k , C2) was unified to L1, which is the live cell tracking ID, but (t k , C3), or it may be possible to unify the IDs with L2, which is the live cell tracking ID assigned to the live cell area in C4). Alternatively, it may be possible to newly generate an ID that is not used as another live cell tracking ID and unify the ID with that.

[0082] (d) to (f) in Fig. 12 show the timing ID of t k-1 The cell image and t k This shows a case where the fusion of live cell regions occurred between the timings when two cell images with timing ID t were taken. kIn the cell image of Fig. 1, the live cell tracking IDs L1 and L2 have already been assigned to the live cell regions with region IDs C2 and C5. k The live cell area in the cell image of the time stamp and the image captured at a later timing (with timing ID t k+1 The regions have already been associated with live cell regions present in the cell images.

[0083] Here, the timing ID is t k The live cell region with region ID C5 in the cell image is focused on, and the timing ID is t k-1 The cell image in Figure 12(e) is the image with timing ID t k The live cell region (t k , C5) and timing ID is t k-1 The live cell regions (t k-1 , C2), (t k-1 , C3). As shown in detail in (e) of Figure 12, (t k , C5) live cell area is (t k-1 , C2), (t k-1 , C3) and the two live cell regions (t k , C5) and the live cell region (t k-1 , C2) overlapping the live cell region is greater than 0. k , C5) and the live cell region (t k-1 , C3) is also greater than 0. In such a case, as shown in (f) of FIG. 12, when (timing ID, region ID) is (t k , C5)(t k-1 , C2)(t k-1 , C3) are considered to be of the same lineage (originating from the same cell region), and are assigned the same live cell tracking ID, L2. If there is a non-viable cell region in the pre-fusion or post-separation cell region, there is no need to perform the above-described method of association.

[0084] 13(a) to (f) show a method similar to the method shown in Fig. 8(a), in which correspondence is first established between cell regions of the same lineage (originating from the same cell region) in cell images taken at previous and subsequent timings for all cell regions included in each cell image. After correspondence is established between cell regions originating from the same cell region present in multiple cell images, correspondence is established between live cell regions originating from the same cell region present in multiple cell images, tracing back in time to the first timing, starting from a live cell region present in a cell image taken at a second timing.

[0085] Here, the timing ID is t k For the cell area present in the cell image of , the image captured at the previous timing (timing ID is t k-1 The timing ID is t k The cell region and timing ID present in the cell image are t k+1 10 shows a step of matching cell regions present in the cell image.

[0086] (a) to (c) in Fig. 13 show the timing ID of t k The cell image and t k+1 This shows a case where cell region separation occurred between the timing when the two cell images were taken. k In the cell image, cell tracking IDs T1 to T5 are assigned to the cell regions with region IDs C1 to C5, respectively, and the cell tracking IDs T1 to T5 are assigned to the cell regions with region IDs C1 to C5, respectively. k-1 The cell regions of the cell images have already been associated with cell regions derived from the same cell region.

[0087] Here, the timing ID is t k The cell region (t k , C4), and the timing ID is t k+1 Figure 13(b) shows the cell image (t k , C4) and the timing ID is t k+1The cell regions ((t k+1 , C2), (t k+1 , C3)) are superimposed. k+1 , C2) and (t k+1 , C3) in any cell region (t k It can be seen that there is an overlap with the cell region of C4).

[0088] On the other hand, (d) to (f) in FIG. 13 show the timing ID t k The cell image and t k+1 This shows a case where fusion of cell regions occurred between the two cell images taken at the same time. k In the cell image, cell tracking IDs T1 to T6 are assigned to the cell regions with region IDs C1 to C6, respectively, and the cell tracking IDs T1 to T6 are assigned to the cell regions with region IDs C1 to C6, respectively. k-1 The cell regions of the cell images have already been associated with cell regions derived from the same cell region.

[0089] Here, the timing ID is t k The cell regions ((t k , C2), (t k , C3)) and timing ID is t k+1 Figure 13(e) shows the cell image (t k , C2), (t k , C3) and timing ID is t k+1 The cell region (t k+1 , C5) are superimposed. k+1 , C5) cell area is (t k , C2), (t k , C3) and overlap with each other.

[0090] As shown in (b) and (e) of Figure 13, if the overlap amount of the cell area between the cell images taken at the previous and next timings is greater than 0, they are considered to be of the same lineage (originating from the same cell area), and are associated so that they have the same cell tracking ID. k , C4), (t k+1 , C2) and (t k+1 , C3) are assigned the same cell tracking ID4. k , C2), (t k , C3) and (t k+1 , C5) are assigned the same cell tracking ID T2. As shown in (e) of Figure 13, cell regions ((t k , C2), (t k , C3)) are determined to be of the same lineage, the cell tracking ID already assigned is updated. Note that the cell tracking ID assigned or updated in the examples of (d) to (f) of Figure 13 may be T3 or a new, unused cell tracking ID. In (c) and (f) of Figure 13, for ease of understanding, cell regions that have the same cell tracking ID and originate from the same cell region are painted with the same pattern. Note that a method for associating cell regions originating from the same cell region that exist in multiple cell images can be achieved by replacing "living cell region" with "cell region" in the calculation of the amount of overlap using the above-mentioned formula (1).

[0091] Using the above method, correspondence between cell regions derived from the same cell region is performed. Note that, in this modified example, a case is shown in which correspondence of the same system is performed for all cell regions in chronological order from the first timing t1, ie, t2 → t3 → ..., but the order of correspondence is not limited to this.

[0092] Finally, similar to the method shown in Fig. 8(a) above, the living cell regions are associated with each other by assigning unique living cell tracking IDs to the living cell regions having the same cell tracking ID as the living cell region at the second timing, going back in time, thereby associating the living cell regions with each other. This allows association of living cell regions originating from the same cell region.

[0093] (Variation 5: Displaying viability assessment results along with image data) It is also preferable to present the determination results obtained in the viability determination step to the user together with image data for observing the progress of the cells. Specific examples of display methods will be described below with reference to Figures 14 and 15. FIG. 14 shows an example of a UI screen displaying a superimposed image of the viability assessment result and a video sequence generated based on cell images taken from the first imaging time t1 to the second imaging time tN. The UI screen 1401 includes an image display screen 1402 displaying the superimposed image of the viability assessment result, a video sequence display screen 1403 displaying the video sequence generated based on cell images taken from the first imaging time t1 to the second imaging time tN, and a UI 1404 for the user to control the display frames of the video sequence. The superimposed image of the viability assessment result may be any image processed so that each cell region present in the cell image taken at the first imaging time t1 can be distinguished as viable or non-viable. For example, an image may be generated in which cell regions determined to be viable and cell regions determined to be non-viable are filled in with different colors. Alternatively, an image may be generated in which lines indicating the outlines of cell regions are drawn on the cell image in different colors or thicknesses for viable and non-viable cells. The video display screen 1403 is a video in which cell images from the first imaging timing t1 to the second imaging timing tN are used as frames. Here, information acquired in the live cell region determination step or the tracking step may be superimposed on each frame image in the video. For example, the cell images at each imaging timing are processed to create a video so that each cell region can be distinguished as live or non-live. Specifically, as with the above-mentioned method, the cell regions may be processed by filling live cell regions and non-live cell regions with different colors, or by drawing contour lines indicating the cell regions with different colors or different thicknesses.

[0094] FIG. 15 shows an example in which cropped images of each cell region at each imaging timing are created as image data for observing the progression of cells, and are displayed side by side based on the imaging timing and the (live) cell tracking ID of the cell region. UI screen 1501 is composed of image display screen 1402, which displays a superimposed image of the viability assessment result, and image display screen 1502, which displays cropped images of each cell region at each imaging timing. The images displayed on image display screen 1402 are the same as those shown in FIG. 14. On image display screen 1502, cropped images of each cell region at each imaging timing are grouped by (live) cell tracking ID and displayed in order of imaging timing. Here, the cropped images are images that are cropped to a size smaller than the captured cell images so that the corresponding cell region is located at the center.

[0095] As described in Modification Example 4, cells grow through fusion and separation, and there may be multiple cell regions assigned the same (live) cell tracking ID at the same imaging timing. In such cases, a cropped image can be created and displayed that includes all cell regions with the same (live) cell tracking ID in cell images captured at the same imaging timing. Alternatively, the correspondence between fusion and separation at previous and subsequent imaging timings may be displayed in the form of a tree diagram. FIG. 16 shows an example of cropped images displayed in the form of a tree diagram. In the image display screen 1602 of the UI screen 1601, cropped images of each cell region at each imaging timing are grouped by (live) cell tracking ID and displayed in order of imaging timing, similar to FIG. 15 . In FIG. 16 , cropped images of cell regions that have undergone fusion or separation are connected by lines, allowing the user to distinguish which cell regions corresponded to which at the previous and subsequent imaging timings.

[0096] 15 and 16 show examples in which cropped images corresponding to each (live) cell tracking ID are arranged, but the displayed cropped images may be switched in response to a user operation. For example, the user may select a cellular region present in a superimposed image of the viability assessment result, and only the cropped images of the cellular region corresponding to the tracking ID assigned to the selected cellular region may be displayed. Furthermore, a UI for narrowing down the imaging timings to be displayed may be further provided, and the imaging timings of the cropped images to be displayed may be narrowed down in response to a user operation.

[0097] As described above, according to this modified example, the user can observe the progress of each cell region in association with the viability assessment result at the first imaging timing. The display method is not limited to the examples in Figures 14, 15, and 16, and may be a combination of the respective display methods or may be switched in response to a user operation. For example, a UI screen may be configured with an image display screen 1402 and a moving image display screen 1403 that display an image on which the viability assessment result in Figure 14 is superimposed, and an image display screen 1502 that displays a cropped image in Figure 15.

[0098] Second Embodiment The program according to this embodiment is an information acquisition program for causing a computer to execute the information acquisition method described above.

[0099] <Third embodiment> The recording medium according to this embodiment stores the information acquisition program in a computer-readable format. [Example]

[0100] Below, an example will be given in which information on cell viability is obtained by the information obtaining method described in this specification. iPS cells were used as the target cells, and cell suspensions that had been previously frozen and thawed were used as cell samples. The freezing rate was set to three levels: -0.5°C / min, -1°C / min, and -2°C / min, and the thawing temperature and time were set to three levels: 4°C and 25 minutes, 25°C and 6 minutes, and 37°C and 80 seconds. By combining these levels, cell samples with high and low viability were prepared.

[0101] The iPS cell suspension obtained in this way was seeded at 13,000 cells per well in a 6-well plate and cultured according to standard methods (https: / / www.cira.kyoto-u.ac.jp / j / research / img / protocol / hiPS_PBMC_protocol.pdf?1732087689398). Time-lapse images of the cells during culture were automatically acquired using an Incucyte S3 (Sartorius). The imaging conditions were 4x magnification, 9 fields of view, phase contrast observation, and a schedule of once per hour. The first image was taken at time 0, and images were taken up to 24 hours later.

[0102] The acquired time-lapse images were subjected to a viable cell region determination process using a self-developed pre-trained classifier, followed by a tracking process and a viability determination process. This allowed us to measure the viability (number of viable cell regions determined to be viable / total number of cell regions) and viable cell count (number of viable cell regions determined to be viable) in the entire field of view at time 0. As a result, the survival rate and number of survivors varied depending on the freezing and thawing conditions, with the ranges being 20% ​​to 46% for the survival rate and 398 to 1017 for the number of survivors.

[0103] <Reference example> As a reference example, the same cell samples as in the Examples were used to measure the viability immediately after cell lysis using a commercially available cell counter, and the cells were seeded and cultured according to a standard method (https: / / www.cira.kyoto-u.ac.jp / j / research / img / protocol / hiPS_PBMC_protocol.pdf?1732087689398), and the confluency was measured on day 4 of culture.

[0104] The viability of cells was measured using a cell counter, either by staining with trypan blue (TB) and counting with a Countess 3 (ThermoFisher) or by staining with acridine orange (AO) and counting with a NucleoCounter NC-200 (Chemometec). Confluency was measured on day 4 of culture by seeding cells at 13,000 cells / well in a 6-well plate, and images were acquired and analyzed using an Incucyte S3 (Sartorius) after culture.

[0105] Figure 17 shows the correlation between the viability in the Example and the confluency on day 4 of culture in the Reference Example, as well as the correlation between the viable cell rate in the Reference Example and the confluency on day 4 of culture in the Reference Example. The viability in the Example correlated with confluency, which is the amount of cells actually obtained by culture (correlation coefficient 0.61). On the other hand, the viable cell rate in the Reference Example had a weak correlation with confluency (both were weak negative correlations). From the above, it was found that the viability in this Example is an index that reflects cell viability better than the viable cell rate currently used as a standard.

[0106] Figure 18 shows the correlation between the viable count in the Example and the confluency on day 4 of culture in the Reference Example. The correlation coefficient was 0.64, which is slightly better than the viability shown in Figure 17. Since it is difficult to strictly maintain a constant cell seeding amount, it is thought that the absolute viable count has a higher correlation with confluency than the viability. Depending on the application, it may be preferable to use the viable count as an index.

[0107] Embodiments of the present invention can also be realized by a computer in a system or apparatus (e.g., an application-specific integrated circuit (ASIC)) that reads and executes computer-executable instructions (e.g., one or more programs) recorded on a storage medium to perform one or more functions of the above-described embodiments, and / or by a computer in a system or apparatus that includes one or more circuits that perform one or more functions of the above-described embodiments, as well as by a method implemented by the computer in the system or apparatus, e.g., by reading and executing the computer-executable instructions from a storage medium to perform one or more functions of the above-described embodiments, and / or by controlling one or more circuits to perform one or more functions of the above-described embodiments. The computer can include one or more processors (e.g., a central processing unit (CPU), a microprocessor unit (MPU)), and can include a separate computer or a network of separate processors to read and execute the computer-executable instructions. The computer-executable instructions may be provided to the computer, for example, from a network or a storage medium. The storage medium may include, for example, one or more of a hard disk, a random access memory (RAM), a read-only memory (ROM), a storage device of a distributed computing system, an optical disk (compact disk (CD), digital versatile disk (DVD), Blu-ray disk (BD), etc.), a flash memory device, a memory card, etc.

[0108] The disclosure of this embodiment includes the following configurations and methods. (Method 1) an image acquisition step of acquiring a plurality of cell images captured at a plurality of different imaging timings including a first timing and a second timing; a live cell region determination step of determining whether a cell region present in each of the plurality of cell images is a live cell region containing a live cell; a tracking step of tracking live cell regions originating from the same cell region by tracing back in time to the first timing, starting from a live cell region present in the cell image captured at the second timing, and associating live cell regions originating from the same cell region present in each of the plurality of cell images captured at the plurality of different imaging timings; An information acquisition method for acquiring information regarding cell viability, characterized by including a viability determination step of determining that a live cell region that was tracked in the tracking step is viable among the cell regions present in the cell image captured at the first timing. (Method 2) The information acquisition method described in Method 1 further includes a viability quantification calculation step of calculating at least one of the number and proportion of live cell areas that are determined to be viable in the viability determination step among the cell areas present in the cell image captured at the first timing. (Method 3) In the image acquisition step, the cell image is a cell image obtained by photographing cells that have been stained with a reagent that stains at least one of live cells and dead cells, 3. The information acquisition method according to Method 1 or 2, wherein in the live cell region determination step, it is determined whether the cell region is a live cell region or not based on the difference in the degree of staining of the cell region by the reagent. (Method 4) An information acquisition method described in any one of methods 1 to 3, characterized in that in the live cell region determination step, the determination is made using a classifier that has been trained to input at least one of at least a portion of the cell image that includes the target cell region and an image feature calculated from at least a portion of the cell image that includes the target cell region, and output the result of determining whether or not the region is a live cell region. (Method 5) An information acquisition method described in any of methods 1 to 4, characterized in that in the tracking process, living cell regions in which fusion or separation occurred between the first timing and the second timing are associated as living cell regions derived from the same cell region. (Configuration 6) An information acquisition program for causing a computer to execute the information acquisition method according to any one of Methods 1 to 5. (Configuration 7) A recording medium storing the information acquisition program according to configuration 6 in a computer-readable format. (Configuration 8) an image acquisition unit that acquires a plurality of cell images captured at a plurality of different imaging timings including a first timing and a second timing; a live cell region determination unit that determines whether a cell region present in each of the plurality of cell images is a live cell region containing a live cell; a tracking unit that tracks live cell regions originating from the same cell region by tracing back in time to the first timing, starting from a live cell region present in the cell image captured at the second timing, and associating live cell regions originating from the same cell region present in each of the plurality of cell images captured at the plurality of different imaging timings; An information acquisition system for acquiring information regarding cell viability, characterized by having a viability determination unit that determines, among the cell regions present in the cell image captured at the first timing, a live cell region that can be tracked by the tracking unit as viable.

[0109] S101 Image acquisition process S102 Viable cell area determination step S103 Tracking process S104 Viability determination step S105 Viability quantitative calculation process 201 Information Acquisition System 202 Image acquisition unit 203 Viable cell area determination section 204 Tracking Department 205 Viability Determination Department 206 Memory section 207 Viability Quantification Calculation Department 208 Data Server 210 Observation Device 211 Input Device 212 Output Device 401 Cell Images 402 Cell area 403 Living cell area 404 Cell regions that were not determined to be viable

Claims

1. an image acquisition step of acquiring a plurality of cell images captured at a plurality of different imaging timings including a first timing and a second timing; a live cell region determination step of determining whether a cell region present in each of the plurality of cell images is a live cell region containing a live cell; a tracking step of tracking live cell regions originating from the same cell region by tracing back in time to the first timing, starting from a live cell region present in the cell image captured at the second timing, and associating live cell regions originating from the same cell region present in each of the plurality of cell images captured at the plurality of different imaging timings; An information acquisition method for acquiring information regarding cell viability, characterized by including a viability determination step of determining that a live cell region that was tracked in the tracking step is viable among the cell regions present in the cell image captured at the first timing.

2. The information acquisition method described in claim 1, further comprising a viability quantification calculation step of calculating at least one of the number and proportion of live cell areas determined to be viable in the viability determination step among the cell areas present in the cell image captured at the first timing.

3. In the image acquisition step, the cell image is a cell image obtained by photographing cells that have been stained with a reagent that stains at least one of live cells and dead cells, 2. The information acquisition method according to claim 1, wherein in the live cell region determination step, it is determined whether the cell region is a live cell region or not based on a difference in the degree of staining of the cell region by the reagent.

4. The information acquisition method described in claim 1, characterized in that in the live cell region determination process, the determination is made using a classifier that has been trained to input at least one of at least a portion of the cell image that includes the target cell region and an image feature calculated from at least a portion of the cell image that includes the target cell region, and output the result of determining whether or not the region is a live cell region.

5. The information acquisition method described in claim 1, characterized in that in the tracking process, living cell regions in which fusion or separation occurred between the first timing and the second timing are associated as living cell regions derived from the same cell region.

6. An information acquisition program for causing a computer to execute the information acquisition method according to any one of claims 1 to 5.

7. 7. A recording medium storing the information acquisition program according to claim 6 in a computer-readable format.

8. an image acquisition unit that acquires a plurality of cell images captured at a plurality of different imaging timings including a first timing and a second timing; a live cell region determination unit that determines whether a cell region present in each of the plurality of cell images is a live cell region containing a live cell; a tracking unit that tracks live cell regions originating from the same cell region by tracing back in time to the first timing, starting from a live cell region present in the cell image captured at the second timing, and associating live cell regions originating from the same cell region present in each of the plurality of cell images captured at the plurality of different imaging timings; An information acquisition system for acquiring information regarding cell viability, characterized by having a viability determination unit that determines, among the cell regions present in the cell image captured at the first timing, a live cell region that can be tracked by the tracking unit as viable.

Images