Staining intensity acquisition method, program, and staining intensity acquisition device

Abstract

In a target region (42) corresponding to at least a portion of a cell, first, a plurality of high-intensity pixels (422) having a relatively high pixel staining intensity, which is a staining intensity derived from a pixel value, are identified. Next, a region staining intensity, which is the staining intensity of the target region (42), is acquired as the average value of the pixel staining intensities of the plurality of high-intensity pixels (422). As a result, the present invention can acquire an appropriate region staining intensity enabling suppression of false negatives and false positives in dyeing determination.

Description

Staining intensity acquisition method, program, and staining intensity acquisition device

[0001] The present invention relates to a technique for acquiring staining intensity of at least a partial region of each cell from a cell image showing a plurality of stained cells. [Reference to Related Applications] This application claims the benefit of priority from Japanese Patent Application JP2023-152127, filed on September 20, 2023, the entire disclosure of which is incorporated herein by reference.

[0002] In recent years, in medical and biological research, single-cell analysis of biological specimens such as tissue and cell specimens has been performed with the aim of elucidating disease mechanisms, biological mechanisms, drug action mechanisms, etc. For example, in multiplex immunostaining analysis, biological substances such as proteins are stained using various methods and quantitative and qualitative analysis is performed by measuring the staining state.

[0003] In Japanese Patent Application Laid-Open No. 2022-123488 (Document 1), a cell region corresponding to a cell, or a region at the center or periphery of a cell, is extracted from an image of stained tissue, and the average density of the entire extracted region or the maximum density within the region is obtained as the "staining density." If the staining density is equal to or greater than a predetermined threshold, the cell is determined to be "positive" for the stain, and if the staining density is equal to or less than the threshold, the cell is determined to be "negative" for the stain. "Positive" means, for example, that the cell has a significant amount of a specific protein or that intracellular organelles are present.

[0004] However, when determining whether a stain is positive or negative based on the average staining intensity (staining density in Literature 1) in a region extracted from a cell image as in Literature 1, there is a risk that a false positive will be determined if the staining density is somewhat high overall, or that a false negative will be determined because the region with high staining density is small. Furthermore, if the determination is based only on the maximum density in the extracted region, the pixel values ​​will be strongly influenced by pixels with abnormalities, and sufficient determination accuracy will not be achieved.

[0005] The present invention aims to obtain appropriate staining intensity in a target region corresponding to at least a portion of each cell from a cell image showing a plurality of stained cells.

[0006] Aspect 1 of the present invention is a staining intensity acquisition method for acquiring, from a cell image showing a plurality of stained cells, a regional staining intensity, which is the staining intensity of a target region corresponding to at least a portion of each cell, comprising the steps of: a) identifying, in the target region, a plurality of high-intensity pixels having relatively high pixel staining intensities, which are staining intensities derived from pixel values; and b) acquiring the regional staining intensities of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

[0007] According to the first aspect of the present invention, it is possible to obtain an appropriate staining intensity of a target region.

[0008] A second aspect of the present invention is the staining intensity acquisition method of the first aspect, wherein the plurality of high-intensity pixels are a predetermined number of pixels in the target region that have the highest pixel staining intensity.

[0009] A third aspect of the present invention is the staining intensity acquisition method of the second aspect, wherein the predetermined number is proportional to the number of pixels included in the target region.

[0010] Aspect 4 of the present invention is the staining intensity acquisition method of Aspect 1 (which may be any one of Aspects 1 to 3), wherein the plurality of high-intensity pixels are spaced apart from each other by a predetermined distance or more.

[0011] Aspect 5 of the present invention is a staining intensity acquisition method according to aspect 1, wherein step a) comprises the steps of: identifying at least one peak pixel in the target region, the peak pixel having a pixel staining intensity higher than the pixel staining intensities of surrounding pixels; and identifying pixels within a circular region of a predetermined size centered on the at least one peak pixel as the plurality of high-intensity pixels.

[0012] Aspect 6 of the present invention is the staining intensity acquisition method of Aspect 1 (which may be any one of Aspects 1 to 5), wherein the cell image represents a tissue specimen.

[0013] A seventh aspect of the present invention is the staining intensity acquisition method according to any one of the first to sixth aspects, wherein the target region is extracted from the cell image as a region showing one entire cell.

[0014] Aspect 8 of the present invention is a staining intensity acquisition method according to any one of aspects 1 to 6, further comprising, prior to step a), steps c) of extracting a provisional target region from the cell image corresponding to at least a portion of each cell, and d) of dividing the provisional target region into an internal region and an outer peripheral region, wherein the target region is either the internal region or the outer peripheral region.

[0015] A ninth aspect of the present invention is a program that causes a computer to acquire a regional staining intensity, which is the staining intensity of a target region corresponding to at least a portion of each cell, from a cell image showing a plurality of stained cells, and execution of the program by the computer causes the computer to perform the following steps: a) identifying a plurality of high-intensity pixels in the target region that have relatively high pixel staining intensities, which are staining intensities derived from pixel values; and b) acquiring the regional staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

[0016] A tenth aspect of the present invention is a staining intensity acquisition device that acquires a regional staining intensity, which is the staining intensity of a target region corresponding to at least a portion of each cell, from a cell image showing a plurality of stained cells, and includes a pixel identification unit that identifies a plurality of high-intensity pixels in the target region that have relatively high pixel staining intensities, which are staining intensities derived from pixel values, and an intensity acquisition unit that acquires the regional staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

[0017] The above and other objects, features, aspects and advantages will become more apparent from the following detailed description of the invention which proceeds with reference to the accompanying drawings.

[0018] 1 is a diagram showing the configuration of a computer that functions as a staining intensity acquisition device. FIG. 2 is a block diagram showing the functional configuration of the staining intensity acquisition device. FIG. 3 is a diagram showing the flow of operation of a computer. FIG. 4 is a diagram showing a cell image. FIG. 5 is a diagram showing a target region. FIG. 6 is a diagram showing positions of high intensity pixels identified in the target region. FIG. 7 is a diagram showing a cell image showing a determination result. FIG. 8 is a diagram showing a state in which the entire target region is non-specifically stained. FIG. 9 is a diagram for explaining another example of operation of the staining intensity acquisition unit. FIG. 10 is a diagram for explaining another example of operation of the staining intensity acquisition unit. FIG. 11 is a diagram for explaining yet another example of operation of the staining intensity acquisition unit. FIG. 12 is a diagram for explaining yet another example of operation of the staining intensity acquisition unit. FIG. 13 is a diagram for explaining another example of operation of the target region acquisition unit. FIG. 14 is a diagram for explaining another example of operation of the target region acquisition unit.

[0019] FIG. 1 is a diagram showing the configuration of a computer 5 functioning as a staining intensity acquisition device according to one embodiment of the present invention. The staining intensity acquisition device is a device that acquires staining intensities of target regions corresponding to at least a portion of each cell from a cell image (hereinafter referred to as "regional staining intensities" to distinguish them from staining intensities for each pixel, which will be described later). A cell image is a digital image showing multiple cells stained with biological substances such as specific proteins or RNA (ribonucleic acid), and is, for example, an image obtained by capturing an image of a biological specimen obtained by an immunostaining method. The cell image may be one of multiple images obtained for each staining in a multiple immunostaining method that repeatedly stains and destains.

[0020] The computer 5 has a typical computer system configuration including a CPU 51, a GPU 52, a ROM 53, a RAM 54, a fixed disk 55, a display 56, an input unit 57, a reading device 58, a communication unit 59, and a bus 50. The CPU 51 performs various arithmetic processing. The GPU 52 performs various arithmetic processing related to image processing at high speed. The ROM 53 stores basic programs. The RAM 54 stores various information. The fixed disk 55 stores information. The display 56 is a display unit that displays various information such as images.

[0021] The input unit 57 includes a keyboard 57a and a mouse 57b for receiving input from an operator. The reader 58 reads information from a computer-readable recording medium 9 such as an optical disk, a magnetic disk, a magneto-optical disk, or a memory card. The display 56, the keyboard 57a, the mouse 57b, and the reader 58 are connected to the bus 50 via an interface (I / F). The communication unit 59 transmits and receives signals to and from devices external to the computer 5.

[0022] The bus 50 is a signal circuit that connects the CPU 51, GPU 52, ROM 53, RAM 54, fixed disk 55, display 56, input unit 57, reading device 58, and communication unit 59. An imaging device 8 external to the computer 5 is connected to the bus 50 via an I / F. The imaging device 8 acquires an image of a specimen as a cell image, and the data is stored in the fixed disk 55.

[0023] In the computer 5, the program 91 is read in advance from the recording medium 9 via the reading device 58 and stored on the fixed disk 55. The program 91 may also be stored on the fixed disk 55 via a network. The CPU 51 and the GPU 52 execute arithmetic processing using the RAM 54 and the fixed disk 55 in accordance with the program 91. The CPU 51 and the GPU 52 function as a calculation unit in the staining intensity acquisition device 1. Other components that function as a calculation unit may be employed in addition to the CPU 51 and the GPU 52.

[0024] 2 is a block diagram showing the functional configuration of the staining intensity acquisition device 1 realized by the computer 5 executing arithmetic processing and the like according to the program 91. FIG. 2 also shows configurations other than the function of the staining intensity acquisition device 1. In FIG. 2, a staining intensity acquisition unit 13 shows the main functional configuration of the staining intensity acquisition device 1. The memory unit 11 and the target region acquisition unit 12 may be included in the functional configuration of the staining intensity acquisition device 1. When the computer 5 further realizes the functional configuration of the staining determination unit 14, the computer 5 becomes a device that determines whether the staining of each cell in a cell image is positive or negative.

[0025] All or part of each functional configuration may be realized by a dedicated electric circuit. Furthermore, these functional configurations may be realized by a plurality of computers. Of the functional configurations shown in FIG. 2 , the storage unit 11 is realized primarily by the RAM 54 and the fixed disk 55. The target region acquisition unit 12, the staining intensity acquisition unit 13, the staining determination unit 14, and the display control unit 15 are realized by the CPU 51, the GPU 52, the ROM 53, the RAM 54, the fixed disk 55, and their peripheral components.

[0026] The storage unit 11 stores cell images, which are images of a biological specimen (hereinafter simply referred to as a "specimen") acquired by the imaging device 8. The stored images are, to be precise, "image data 81" as shown in FIG. 2, and in the following description, processing of the images refers, to be precise, to processing of the image data 81.

[0027] Figure 3 is a diagram showing the flow of operation of the computer 5, and shows the operation of obtaining the regional staining intensity (which may be considered as the staining intensity of each cell), which is the staining intensity of each target region in a cell image, and determining whether the staining of the target region is positive or negative.

[0028] First, a specimen is imaged and cell images are prepared in the storage unit 11 (step S11) by the control of the imaging device 8 by the computer 5 or by the operation of the imaging device 8. As illustrated in Fig. 4, the cell images 4 stored in the storage unit 11 show a plurality of cells 41. In Fig. 4, each cell 41 is simply represented by a circle.

[0029] Next, the target region acquisition unit 12 acquires, as a target region, a region corresponding to at least a portion of each cell 41 from the cell image 4 (step S12). If the cell image 4 to be processed is one of multiple images obtained for each staining in a multiple immunostaining method that repeatedly stains and destains, a region corresponding to the target region may be acquired from a cell image stained differently from the cell image 4 (i.e., a differently stained image of the same specimen), and this region may be used as the target region of the cell image 4. The function of the region division unit 121 in the target region acquisition unit 12 will be described later. Hereinafter, the multiple target regions acquired by the target region acquisition unit 12 will be referred to as a "target region group."

[0030] A target region may be a so-called "cell region" corresponding to an entire cell 41, or may be a portion of a cell region, such as a region corresponding to the nucleus of a cell. Various known methods can be used to acquire the target region. For example, various known methods for acquiring a cell region, various known methods for acquiring a nuclear region, and various known methods for acquiring at least a portion of a cell region that is not bound to the cellular tissue can be used. When single-cell analysis is performed, the target region and the cell preferably have a one-to-one correspondence. That is, the target region is at least a portion of the cell region, and the cell region and the target region have a one-to-one correspondence. In this way, the target region is a region corresponding to at least a portion of each cell. The computer 5 does not necessarily have a function for extracting a cell region, as long as it has a function for extracting a target region.

[0031] Next, the staining intensity acquisition unit 13 selects a target region to be the first processing target (step S13). Fig. 5 is a diagram illustrating an example of the target region 42 when the target region 42 is a cellular region. A stained region 421 exists within the target region 42. In Fig. 5, the region 421 is shown as a circle with hatched lines to facilitate understanding, but in reality, the region 421 has varying shades of gray and various shapes, and some regions have unclear contours.

[0032] The staining intensity acquisition unit 13 includes a pixel identification unit 131 and an intensity acquisition unit 132. The pixel identification unit 131 identifies, among the pixels in the target region 42, a plurality of pixels (hereinafter referred to as "high-intensity pixels") having a relatively high staining intensity (hereinafter referred to as "pixel staining intensity") derived from the pixel value (step S14). The "staining intensity" in "region staining intensity" and "pixel staining intensity" refers to the intensity at which the stain appears in a cell image. For example, in an image of stained color captured using a bright-field imaging system, if a strongly stained region appears dark, the darker the image, the higher the staining intensity. Conversely, in an image of fluorescent dye staining, if a strongly stained region appears bright, the brighter the image, the higher the staining intensity.

[0033] Therefore, when the cell image is a monochrome image, the larger the pixel value, the higher or lower the pixel stain intensity. When the cell image is a color image, the larger the value of a specific color element of a pixel, or the value calculated from multiple color elements, the higher or lower the pixel stain intensity. Note that when the cell image is a color image, an image that has previously undergone processing such as color separation or color extraction to extract only the expected stain color components may be treated as the new cell image to be processed. Even when the cell image is a monochrome image, an image that has undergone image processing may be treated as the new cell image. As described below, the same applies to the relationship between the level of regional stain intensity derived from the pixel values ​​of multiple high-intensity pixels and the brightness and color of the target region 42 within the image.

[0034] Generally speaking, "stain intensity" is a value derived from the pixel values ​​of a single pixel or multiple pixels contained in a region (the pixel stain intensity may be the pixel value itself) and indicates the intensity of staining. More precisely, "pixel stain intensity" is a value based on the pixel value of a single pixel, and "region stain intensity" is a value based on the pixel values ​​of multiple high-intensity pixels contained in the region of interest 42.

[0035] As described above, an increase in pixel stain intensity results in a simple increase or decrease in pixel value (including cases where the pixel value corresponds to the value of one color component of a color pixel value or a value derived from multiple color components). Therefore, the pixel identification unit 131 of the stain intensity acquisition unit 13 can identify multiple high-intensity pixels in the target region 42 that have relatively high pixel stain intensities without actually calculating pixel stain intensities. For example, the pixel identification unit 131 identifies a predetermined number of pixels with the highest pixel value or a predetermined number of pixels with the lowest pixel value as high-intensity pixels. This allows the predetermined number of pixels with the highest pixel stain intensities in the target region 42 to be easily identified as multiple high-intensity pixels. The "predetermined number of pixels with the highest pixel stain intensities" refers to the predetermined number of pixels ranked highest when the pixels are sorted in descending order of pixel stain intensity.

[0036] FIG. 6 is a diagram illustrating the positions of four high-intensity pixels 422 identified in a target region 42 using "+" marks. The number of high-intensity pixels 422 identified in each target region 42 is not particularly limited. In a preferred example, the number of high-intensity pixels 422 is proportional to the number of pixels included in each target region 42 in an image. In another preferred example, the number of high-intensity pixels 422 is constant and does not depend on the size of the target region 42. The ratio of the number of high-intensity pixels 422 to the number of pixels in the target region 42 and the number of high-intensity pixels 422 are appropriately set depending on the size of the actual target region 42, the resolution of the cell image, and the staining characteristics (e.g., the amount of target protein or intracellular organelle present, the size of the target, and the ease of staining).

[0037] The ratio of the number of high-intensity pixels 422 to the number of pixels in the target region 42 may be automatically determined by learning. For example, for cell images with various stains, a computer learning device may be made to learn combinations of whether multiple target regions 42 are positive or negative and the optimal number of high-intensity pixels 422, and when an input image is input to the learning device, the ratio of the number of high-intensity pixels 422 to the number of pixels in the target region 42 may be output.

[0038] Next, the intensity acquisition unit 132 acquires the average value of the pixel stain intensities of the identified multiple high-intensity pixels 422 as the regional stain intensity of the target region 42 (step S15). As described above, pixel values ​​can be used instead of pixel stain intensities, so in step S15, the average value of the pixel values ​​of the high-intensity pixels 422 may be acquired as the regional stain intensity. Of course, the pixel stain intensity of each high-intensity pixel 422 may be calculated, and the average value of the pixel stain intensities may be acquired as the regional stain intensity of the target region 42. In this way, the intensity acquisition unit 132 acquires the regional stain intensity of the target region 42 based on the pixel values ​​of the multiple high-intensity pixels 422. If the number of high-intensity pixels 422 is proportional to the number of pixels included in the target region 42, the regional stain intensity can be acquired while suppressing the influence of the size of the target region 42.

[0039] Once the regional staining intensity for one target region 42 has been obtained, staining intensity acquisition unit 13 checks whether the next target region 42 for which the regional staining intensity has not been determined exists within the target region group (step S16), and if so, selects the next target region 42 and acquires its regional staining intensity (steps S13 to S15). Once the regional staining intensities for all target regions 42 have been obtained (step S16), the process proceeds to processing by staining determination unit 14.

[0040] In the staining determination unit 14, a threshold value is set in advance by the operator via the input unit 57, and the regional staining intensity of each target region 42 in the target region group is compared with the threshold value. If the regional staining intensity of a target region 42 is higher than the threshold value, the target region 42 is determined to be positive for staining (step S17). Note that the smaller the average pixel value of the high-intensity pixels included in the target region 42, the higher the regional staining intensity. If the average pixel value is directly compared with the threshold value in the staining determination unit 14, the target region 42 is determined to be positive for staining if the average pixel value is smaller than the threshold value. The determination result 82 for each target region 42 is stored in the memory unit 11.

[0041] The determination result 82 is displayed on the display 56 under the control of the display control unit 15. For example, as shown in Figure 7, among the target areas 42 of the cell image 4, the positive target areas 42a are colored and the cell image 4 is displayed on the display 56.

[0042] The staining intensity acquisition device 1 can appropriately determine whether a stain is positive or negative by utilizing the pixel staining intensity of the high-intensity pixels 422. For example, as shown in FIG. 8 , the entire target region 42 may be nonspecifically stained, with the entire target region 42 exhibiting approximately 60% of its maximum staining intensity. Even if such a staining condition should not be determined as positive, if the average value of the pixel staining intensity of the entire target region 42 is acquired as the regional staining intensity of the target region 42, as in the past, a threshold value of 50% would result in the target region 42 being erroneously determined as positive. On the other hand, in the case of FIG. 6 , even though a positive determination should be made, because there are few regions with high staining intensity, if the average pixel staining intensity of the entire target region 42 is calculated as the regional staining intensity, and this value were 40%, the target region 42 would be erroneously determined as negative.

[0043] In contrast, when the pixel staining intensity of the high-intensity pixels 422 is used, and the average of the pixel staining intensities of the high-intensity pixels 422 is obtained as the regional staining intensity, the regional staining intensity is obtained as, for example, 80%, and the target region 42 is determined to be positive. In this way, by using the pixel staining intensity of the high-intensity pixels 422 (i.e., by using the pixel values), it is possible to appropriately determine a region as positive when it has a strongly stained region, thereby reducing false positive and false negative determinations.

[0044] Stain determination using the high-intensity pixels 422 is particularly suitable for stain determination of cell images stained with nuclear or cytoplasmic staining, in which only a portion of the target region 42 is deeply stained. Appropriate stain determination can also be performed even when there is noise called background, in which part or all of the cell tissue is lightly and nonspecifically colored. Because background is more likely to occur when staining tissue specimens than cultured cells, stain determination using the high-intensity pixels 422 is suitable for stain determination of tissue specimens.

[0045] Furthermore, staining determination using the high-intensity pixels 422 is suitable for staining determination of proteins that are highly localized or unevenly distributed within cells in the presence of a background, and is more suitable for staining determination of fibrillarin, which is localized in the nucleolus, than for staining determination of actin, which forms the cytoskeleton.Furthermore, it is also suitable for staining determination of images with a low S / N ratio, and is more suitable for staining determination using an enzyme antibody technique that uses a coloring agent and bright-field optical system, than for staining determination using a fluorescent dye and fluorescent optical system.

[0046] All of the effects of determining regional staining intensity using the high intensity pixel 422 described above are also applicable to determining regional staining intensity by identifying high intensity pixels using peak pixels, as described below.

[0047] 9 and 10 are diagrams illustrating another example of the operation of the staining intensity acquisition unit 13. FIG. 9 is a diagram illustrating an example of a target region 42 that should be determined to be negative. In the target region 42 in FIG. 9, there is one strongly stained region 421a and multiple small, weakly stained regions 421b that should not be recognized as stained. When high-intensity pixels 422 are identified in the example of FIG. 9, if a predetermined number of pixels with high pixel staining intensities are extracted from region 421a, the average pixel staining intensity of the high-intensity pixels 422 will be high, and the target region 42 will be erroneously determined to be positive.

[0048] Therefore, the improved staining intensity acquisition unit 13 acquires multiple high-intensity pixels 422 as pixels spaced apart from each other by a predetermined distance or more. Specifically, as shown in FIG. 10 , first, pixel 422a, which has the highest pixel staining intensity, is identified as the high-intensity pixel 422. Next, pixel 422b, which has the highest pixel staining intensity among the pixels located at least the radius of the dashed line 43 from pixel 422a, is identified as the next high-intensity pixel 422. Next, pixel 422c, which has the highest pixel staining intensity among the pixels located at least the radius of the dashed line 43 from all of the previously identified high-intensity pixels 422a and 422b, is identified as the next high-intensity pixel 422. By repeating the above process, a predetermined number of high-intensity pixels 422 (pixels 422a to 422d shown in FIG. 10 ) are identified.

[0049] Because the pixel staining intensities of pixels 422b, 422c, and 422d are not high, the regional staining intensity, which is the average value of the pixel staining intensities of high-intensity pixels 422, is low. As a result, the target region 42 is correctly determined to be negative. By spacing multiple high-intensity pixels 422 apart, the regional staining intensity can be determined while suppressing the strong influence of specific regions with high staining intensities.

[0050] The distance by which the multiple high-intensity pixels 422 should be separated, i.e., the radius of the circle indicated by the dashed line 43 in Fig. 10, may be determined automatically. For example, the distance may be determined as a value proportional to the average radius of the multiple cellular regions when viewed as a circle, and inversely proportional to the square root of the number of high-intensity pixels 422 to be identified. In this way, the "predetermined distance" by which the multiple high-intensity pixels 422 should be separated may differ for each target region 42.

[0051] Next, another example of the operation of staining intensity acquisition unit 13 in Fig. 2 will be described with reference to Fig. 11 and Fig. 12. In this example of the operation, the operation shown in Fig. 11 is executed by pixel identification unit 131 in step S14 in Fig. 3.

[0052] First, the pixel identification unit 131 identifies at least one peak pixel 423 in the target region 42, as shown in FIG. 12 (step S141). The peak pixel 423 is a pixel in the target region 42 whose pixel stain intensity is higher than that of surrounding pixels, i.e., a pixel at a position where the pixel stain intensity is at its peak when viewed two-dimensionally. However, to exclude peak pixels in regions with low stain intensity, a lower limit is set for the pixel stain intensity required to be identified as a peak pixel 423. Furthermore, even if a pixel exists at a peak position, if the number of surrounding pixels that form the peak is small, the pixel may be excluded from the list of peak pixels. A peak pixel is a pixel that has a pixel stain intensity equal to or greater than a predetermined intensity and is located at the apex of a peak (of stain intensity) equal to or greater than a predetermined magnitude.

[0053] Next, the pixel identifying unit 131 identifies, as a plurality of high-intensity pixels, pixels within a circular region 424 of a predetermined size centered around at least one peak pixel 423 (step S142). The radii of the circular regions 424 may all be the same, or may be larger as the pixel staining intensity of the peak pixel 423 increases. In other words, the "predetermined size" of the circular regions 424 may all be the same, or may differ for each circular region 424.

[0054] Thereafter, intensity acquisition unit 132 acquires the average value of the pixel dye intensities of all pixels within at least one circular region 424 as the regional dye intensity of target region 42. If the number of circular regions 424 is two or more, the average value of the pixel dye intensities in each circular region 424 may be calculated, and then the average value of the plurality of average values ​​may be used as the regional dye intensity of target region 42. In this manner, intensity acquisition unit 132 acquires the regional dye intensity of target region 42 based on the pixel dye intensities of high-intensity pixels within circular region 424 of a predetermined size centered on at least one peak pixel 423.

[0055] Identification of the peak pixel 423 can be easily performed using, for example, the open-source computer vision library "Open CV." By acquiring the regional staining intensity of the target region 42 using the peak pixel 423, more accurate staining determination can be performed when staining occurs in which multiple positive spots are scattered within the target region 42. Therefore, acquiring the regional staining intensity using the peak pixel 423 is suitable when multiple peak pixels 423 are identified for one target region 42 by the pixel identification unit 131. Staining determination using the peak pixel 423 is particularly suitable for staining determination of a cell image stained with nuclear or cytoplasmic staining, in which only a portion of the target region 42 is deeply stained.

[0056] Next, another example of the operation of the target region acquisition unit 12 of FIG. 2 will be described with reference to FIGS. 13 to 15. In this example, steps S21 and S22 of FIG. 13 are executed by the target region acquisition unit 12 instead of step S12 of FIG. 3. The target region acquisition unit 12 first extracts a group of provisional target regions from the cell image 4 (step S21). A provisional target region may be a so-called "cell region" corresponding to an entire cell, or may be a portion of a cell region, such as a region corresponding to a cell nucleus. When single-cell analysis is performed, preferably, there is a one-to-one correspondence between the provisional target region and the cell. That is, the provisional target region is at least a portion of the cell region, and there is a one-to-one correspondence between the cell region and the provisional target region. In this way, the provisional target region is a region corresponding to at least a portion of each cell. The computer 5 does not necessarily have a function to extract a cell region, as long as it has a function to extract a provisional target region.

[0057] FIG. 14 is a diagram illustrating a provisional target region 44. The region 421, indicated by a hatched circle, represents a dyed region, but in reality, the region 421 has various sizes and shapes, and the region 421 has various shades of gray. Some regions 421 have unclear contours. The target region acquisition unit 12 has the function of the region division unit 121 shown in FIG. 2 , and the region division unit 121 divides each provisional target region 44 into an inner region 441 on the inside and an outer region 442 on the periphery (step S22). The outer region 442 is a region of a fixed width that follows the outer edge of the provisional target region 44. The inner region 441 is a region further inside than the outer region 442. The target region acquisition unit 12 then acquires either the inner region 441 or the outer region 442 as the target region. That is, the inner regions 441 of all provisional target regions 44 are acquired as target regions, or the outer regions 442 of all provisional target regions 44 are acquired as target regions. Thereafter, steps S13 to S17 shown in FIG.

[0058] That is, when the interior region 441 is treated as the target region, pixels in the interior region 441 with relatively high pixel staining intensities are identified as high-intensity pixels 422 in step S14 in FIG. 3 . In FIG. 14 , high-intensity pixels 422 are illustrated by a "+" mark. On the other hand, when the outer peripheral region 442 is treated as the target region, pixels in the outer peripheral region 442 with high pixel staining intensities are identified as high-intensity pixels 422 in step S14, as illustrated in FIG. 15 . Then, the regional staining intensities of the target regions are obtained (step S15). After the regional staining intensities of all target regions are obtained (step S16), staining determination is performed for all target regions (step S17).

[0059] By using the inner region 441 as the target region, it is possible to appropriately determine the staining of the inner region 441 even if staining occurs non-specifically with respect to the outer region 442. For example, when it is desired to stain and detect a specific protein in the nucleus of a cell or its vicinity, it is possible to eliminate the influence of staining of the cell membrane. It is also possible to eliminate the influence of staining of adjacent cells. When the outer region 442 is used as the target region, it is possible to appropriately determine the staining of the outer region 442 even if staining occurs non-specifically with respect to the inner region 441. For example, when it is desired to stain and detect a specific protein in the cell membrane, it is possible to eliminate the influence of staining of anything other than the cell membrane.

[0060] Of course, the provisional target region 44 is not limited to a cellular region (a region showing the entire cell), and for example, the provisional target region 44 may be extracted as the nucleus and its surrounding region, with the region inside the nucleus acquired as the inner region 441 and the region on the periphery of the nucleus acquired as the outer region 442. By dividing the provisional target region 44 into the inner region 441 and the outer region 442 and using only one of them as the target region, artifacts in the analysis can be effectively suppressed, and more robust regional staining intensities can be acquired.

[0061] The above embodiment can be modified in various ways.

[0062] In the example shown in FIG. 6 , a predetermined number of pixels in the target region 42 with the highest pixel staining intensities are identified as the plurality of high-intensity pixels 422. In the example shown in FIG. 12 , at least one peak pixel 423 and its surrounding pixels are identified as the plurality of high-intensity pixels. Thus, high-intensity pixels may be identified by various methods as long as they are pixels with relatively high pixel staining intensities. High-intensity pixels are preferably included within the top 50% (more preferably, the top 30%) of all pixels in the target region 42 with high pixel staining intensities. In other words, high-intensity pixels may be selected from the top 50% of pixels with high pixel staining intensities using various techniques. Then, various calculations are performed on the pixel values ​​of the plurality of high-intensity pixels to obtain the regional staining intensities of the target region 42.

[0063] 3 may be modified in various ways within the scope of similar effects. For example, each time the regional staining intensity of one target region 42 is acquired, the regional staining intensity may be compared with a threshold value, and the process may proceed to acquiring the regional staining intensity of the next target region 42.

[0064] The method of using either the inner region 441 or the outer region 442 in FIGS. 13 to 15 can also be applied to the method of obtaining the regional staining intensity using the peak pixel 423 shown in FIGS. 11 and 12.

[0065] The staining in this embodiment is not limited to multiple staining, nor is it limited to staining using antibodies. This embodiment can be used to determine the staining of various stains that stain biological substances. This embodiment is preferably suitable for cases where staining results in shading of cells. It is also suitable for staining specific proteins.

[0066] The configurations in the above-described embodiment and each modification may be combined as appropriate as long as they are not mutually contradictory.

[0067] While the invention has been particularly illustrated and described, it should be understood that the foregoing description is illustrative and not restrictive, and that numerous modifications and variations are possible without departing from the scope of the invention.

[0068] REFERENCE SIGNS LIST 1 stain intensity acquisition device 4 cell image 5 computer 13 stain intensity acquisition unit 41 cell 42 target region 44 provisional target region 91 program 131 pixel identification unit 132 intensity acquisition unit 422 high intensity pixel 423 peak pixel 424 circular region (centered on peak pixel) 441 inner region 442 outer peripheral region S14, S15, S21, S22, S141, S142 steps

Claims

1. A staining intensity acquisition method for acquiring regional staining intensity, which is the staining intensity of a target region corresponding to at least a portion of each cell, from a cell image showing a plurality of stained cells, comprising the steps of: a) identifying a plurality of high-intensity pixels in the target region, the pixel staining intensity being a staining intensity derived from pixel values, which is relatively high; and b) acquiring the regional staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

2. A staining intensity acquisition method according to claim 1, wherein the plurality of high intensity pixels is a predetermined number of pixels in the target region having the highest pixel staining intensity.

3. A staining intensity acquisition method according to claim 2, wherein the predetermined number is proportional to the number of pixels contained in the target region.

4. A staining intensity acquisition method as claimed in claim 1, wherein the plurality of high intensity pixels are spaced apart from each other by a predetermined distance or more.

5. A staining intensity acquisition method as described in claim 1, wherein step a) comprises the steps of: identifying at least one peak pixel in the target region, the pixel staining intensity of which is higher than the pixel staining intensities of surrounding pixels; and identifying pixels within a circular region of a predetermined size centered on the at least one peak pixel as the plurality of high intensity pixels.

6. A staining intensity acquisition method according to claim 1, wherein the cell image represents a tissue specimen.

7. A staining intensity acquisition method according to any one of claims 1 to 6, wherein the target region is extracted from the cell image as a region showing an entire cell.

8. A staining intensity acquisition method as described in any one of claims 1 to 6, further comprising, prior to step a), the steps of c) extracting a provisional target region from the cell image corresponding to at least a portion of each of the cells, and d) dividing the provisional target region into an internal region and an outer peripheral region, wherein the target region is either one of the internal region or the outer peripheral region.

9. A program that causes a computer to obtain regional staining intensity, which is the staining intensity of a target region corresponding to at least a portion of each cell, from a cell image showing a plurality of stained cells, wherein the execution of the program by the computer causes the computer to execute the following steps: a) identifying a plurality of high-intensity pixels in the target region, which have relatively high pixel staining intensity, which is a staining intensity derived from pixel values; and b) obtaining the regional staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

10. A staining intensity acquisition device that acquires a regional staining intensity, which is the staining intensity of a target region corresponding to at least a portion of each cell, from a cell image showing a plurality of stained cells, comprising: a pixel identification unit that identifies a plurality of high-intensity pixels in the target region, the pixel staining intensity being a staining intensity derived from pixel values, and an intensity acquisition unit that acquires the regional staining intensity of the target region based on the pixel values ​​of the plurality of high-intensity pixels.

Images