JP2001059842A - Pathological diagnostic apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform a pathological diagnosis of high accuracy and a pathological diagnosis of a kind which requires the judgment of a physician or the like in conventional cases, by a method wherein information such as information on the overlap of a cell nucleus inside the image of a living-body tissue, information on the distribution state of a cave or information on a positional relationship between the cave and the cell nucleus in a closest position in its circumference is quatitized properly so that the information can be used for a pathological diagnosis by a pathological diagnostic apparatus. SOLUTION: This pathological diagnositic apparatus is provided with a pathological diagnostic feature judgment means 50, by which a feature by the positional relationship and the distribution of a cell nucleus or a cave used for a pathological disgnosis is judged as a numerical value on the basis of the image of a living-body tissue. The pathological diagnosistic feature judgment means is provided with a cell nucleus feature judgment means 51 by which the overlap degree of the cell nucleus as a whole is judged as a cell-nucleus structural feature. The judgment means is provided with a cave-feature judgment means 52 by which the distribution state of the cave is judged as a cave structural feature. The judgment means is provided with a cell nucleus-cave mutual-feature judgment means 53 by which a positional relationship between the cave and a nearest cell nucleus group constituted of a plurality of cell nuclei is judged as a cell nucleus-cave mutual structural feature.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pathological diagnosis apparatus, and more particularly to a pathological diagnosis apparatus for performing a pathological diagnosis by extracting features used for diagnosis by image processing from an image of a living tissue.

[0002]

2. Description of the Related Art A specimen is prepared from a living body tissue such as a human body and is imaged by a photographing apparatus. Diagnosis of benign and malignant tumors is performed based on tissue structure characteristics defined by cell nuclei and cavities in the image. A device for performing a pathological diagnosis based on this image, which is called a pathological diagnosis, is provided as a pathological diagnostic device.
Here, the cavity means a non-structural region existing in a gap between cell nuclei in a cell nucleus region observed in an image.

[0003] As in the pathological diagnosis, there is a cytology as a method for judging whether a tumor is benign or malignant. In the cytology, a diagnosis is made by paying attention to the texture of each cell. on the other hand,
Since pathological diagnosis is characterized by focusing on structures such as the shape and distribution of many cell nuclei and cavities, the diagnostic accuracy of pathological diagnosis is higher than that of cytodiagnosis.

Since oversight of malignancy is never allowed in diagnosis and accurate judgment is required, a pathological diagnosis apparatus requires a quantification process that shows a clear difference between benign and malignant in recognizing the characteristics of the biological tissue structure. is there.

[0005] Examples of conventional devices for diagnosing benign and malignant living tissues are disclosed in Japanese Patent Application Laid-Open Nos.
223868.

[0006] The cytodiagnosis apparatus disclosed in Japanese Patent Application Laid-Open No. 57-15367 measures morphological characteristics such as nucleus area and nucleus concentration of individual cells in an image, and determines the degree of malignancy of the cells based on the measured values. It is to judge.

The cytodiagnosis apparatus disclosed in Japanese Patent Application Laid-Open No. 58-223868 extracts a cell nucleus from an isolated cell region in an image and, at the same time, extracts a cytoplasm from a cell mass region in parallel, thereby performing cytodiagnosis at high speed and in detail. It is what you do.

Japanese Patent Application Laid-Open No. H10-197522 discloses an example of an apparatus for examining information on a tissue in a biological tissue image in addition to examining individual cells as in these cell diagnostic apparatuses. This pathological tissue diagnosis support apparatus performs measurement and measurement of the number and area of cell nuclei, and determines the degree of conformity of a biological tissue image to a plurality of predetermined diagnostic categories representing histopathological features.

[0009]

As described above, in the conventional pathological diagnosis apparatus, diagnosis is performed only on individual cells, or information on a tissue in a biological tissue image disclosed in Japanese Patent Application Laid-Open No. 10-197522 is disclosed. In the device for examining the number and area of cell nuclei, which can be easily compared by numerical values, while examining information such as the positional relationship between cell nuclei, the distribution of cavities, and the positional relationship between cavities and surrounding cell nuclei, etc. However, features useful in pathological diagnosis are excluded from the investigation.

[0010] As described above, information on the positional relationship and distribution state of cell nuclei and cavities is conventionally useful for pathological diagnosis because it requires a special quantification technique of expressing appropriate numerical values and the like. Nevertheless, the diagnosis could not be made automatically by the pathological diagnosis device.

Further, in the conventional pathological diagnosis apparatus, it was not possible to make a determination on the information on the positional relationship and distribution state of cell nuclei and cavities as information useful in pathological diagnosis as described below.

First, there is information on the degree of overlap of cell nuclei.
In pathological diagnosis, a benign diagnosis is made when cell nuclei overlap significantly. The case where the pathologist determines that the cell nuclei overlap significantly is when the number of cell nuclei in the region of interest is large, and when the distance between the cell nuclei is small.

Second, there is information on the bias of the distribution of the positions of the cavities. In pathological diagnosis, when the cavities are evenly distributed in the region to be diagnosed, it is diagnosed as malignant. That is, if the cavities are uniformly distributed, it is malignant, and if the distribution of the cavities is uneven, it is benign.

Third, there is information on the positional relationship between the cavity and the cell nucleus located closest to the cavity. In pathological diagnosis, when there is a certain large distance between a cavity and a cell nucleus around the cavity, it is diagnosed as malignant.

A first object of the present invention is to solve the above-mentioned drawbacks of the prior art and to appropriately quantify information such as the positional relationship and distribution state of cell nuclei and cavities in a living tissue image to obtain such information. Can be used for pathological diagnosis by a pathological diagnosis device, and provides a pathological diagnosis device that conventionally performs a type of pathological diagnosis that requires the judgment of a doctor or the like and cannot be performed by a pathological diagnostic device, or that performs a highly accurate pathological diagnosis. Is to do.

A second object of the present invention is to solve the above-mentioned drawbacks of the prior art, appropriately quantify the information on the overlap of cell nuclei in a biological tissue image, and perform a pathological diagnosis based on the information on the overlap of cell nuclei. It is to provide a diagnostic device.

A third object of the present invention is to solve the above-mentioned drawbacks of the prior art, appropriately quantify information on the distribution state of a cavity in a living tissue image, and perform pathological diagnosis based on the information on the distribution state of the cavity. An object of the present invention is to provide a pathological diagnosis apparatus for performing the diagnosis.

A fourth object of the present invention is to solve the above-mentioned drawbacks of the prior art and appropriately quantify information on the positional relationship between a cavity in a biological tissue image and the nearest cell nucleus around the cavity. An object of the present invention is to provide a pathological diagnosis apparatus for performing a pathological diagnosis based on information on a positional relationship between the cavity and a cell nucleus.

[0019]

In order to achieve the above object, a pathological diagnosis apparatus according to the present invention comprises a pathological diagnosis characteristic determining means for determining characteristics used for pathological diagnosis from an image of a living tissue,
The pathological diagnostic feature of the pathological diagnostic apparatus of the present invention according to claim 2, wherein the pathological diagnostic feature determination means determines a feature in the image based on a positional relationship or distribution state of cell nuclei or cavities. The determining means includes a cell nucleus feature determining means for determining the degree of overlap of the entire cell nucleus as the cell nucleus structural feature.

According to a third aspect of the present invention, the cell nucleus characteristic determining means of the pathological diagnosis apparatus obtains an area occupied by all cell nuclei in the image, and determines the degree of overlap using the area.

According to a fourth aspect of the present invention, the cell nucleus feature determining means of the pathological diagnosis apparatus obtains an interval between centers of gravity of the cell nuclei of adjacent cell nuclei in the image, and determines the degree of overlap using the interval. It is characterized by determining.

According to a fifth aspect of the present invention, in the pathological diagnosis apparatus of the present invention, the pathological diagnostic feature determining means includes a cavity characteristic determining means for determining a distribution state of the cavity as a cavity structure characteristic.

In the pathological diagnosis apparatus according to a sixth aspect of the present invention, the cavity characteristic determining means enlarges a portion indicating the cavity in the image with respect to the image, and determines a portion indicating the cavity in the image after the enlargement. With the occupied area, a cavity shape enlarging process for expressing information on the distribution state of the cavity in the image before the enlargement is performed, and an area occupied by a portion indicating the cavity in the image after the cavity shape enlarging process is obtained. It is characterized in that the distribution state is determined using the above.

In the pathological diagnosis apparatus according to the present invention, the cavity shape enlarging process performed by the cavity feature determining means may include a step of determining whether or not the image is in contact with a boundary of an enlarging process target range and with another cavity. Performs a selection process of selecting all the cavities in the image that are equal to or larger than the determined threshold as cavities to be subjected to enlargement processing. , An image rewriting process is performed to newly include pixels within a predetermined range in the interior of the cavity, and an image obtained thereby is again subjected to the above-described selection process,
Until there are no more cavities to be selected by the first-stage selection process, the series of processes of the selection process and the first-time image rewriting process are repeated to enlarge the cavity shape.

In the pathological diagnosis apparatus according to the present invention, the cavity characteristic determining means obtains a distance between the centers of gravity of the individual cavities in the image, and determines the distribution state using the distance between the centers of gravity. It is characterized by determining.

According to a ninth aspect of the present invention, in the pathological diagnosis apparatus according to the present invention, the pathological diagnosis feature judging means includes a cell nucleus-cavity mutual structural feature, the nearest cell nucleus constituted by the cavity and a plurality of cell nuclei existing around the cavity. It is characterized by comprising a cell nucleus-cavity mutual feature determining means for determining a positional relationship with a group.

According to a tenth aspect of the present invention, in the pathological diagnosis apparatus according to the present invention, the cell nucleus-cavity mutual characteristic determining means obtains a length of a circumscribed rectangle of the outline of the cavity and a length of a circumscribed rectangle of the group of nearest cell nuclei. The positional relationship between the cavity and the nearest cell nucleus group is determined using a contour circumscribed rectangle length and the nearest cell nucleus group circumscribed rectangle length.

[0028] The cell nucleus-cavity mutual feature determination means of the pathological diagnosis apparatus of the present invention according to claim 11 is characterized in that: the area of the cavity, the area of each cell nucleus of the group of the closest cell nuclei, and the individual area of the group of the closest cell nuclei. Determining the area of a closed region connecting the centers of gravity of the cell nuclei, and using the area of the cavity, the area of the cell nucleus, and the area of the closed region to determine the positional relationship between the cavity and the nearest cell nucleus group. It is characterized by.

[0029] In the pathological diagnosis apparatus according to the twelfth aspect of the present invention, the cell nucleus-cavity mutual feature determining means obtains a distance between the contour of the cavity and each cell of the nearest cell nucleus group, and uses this distance. And determining the positional relationship between the cavity and the nearest cell nucleus group.

[0030]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing a configuration of a pathological diagnosis apparatus according to the first embodiment of the present invention. In the present embodiment, a pathological diagnosis is performed based on the overlapping of cell nuclei.

The feature of the cell nucleus structure in pathological diagnosis is to examine the degree of overlap of cell nuclei, etc., and when the overlap of cell nuclei is remarkable, it is diagnosed as benign.

If the cell nuclei overlap significantly in the image,
The area of the cell nucleus occupies a large area in the area under investigation.

For this reason, the pathological diagnosis apparatus according to the present embodiment quantifies the degree of overlap of the cell nuclei numerically by taking the ratio of the total area of the cell nucleus area to the total area of the non-cell nucleus area, and based on the degree of overlap of the cell nuclei. Performing Pathological Diagnosis Referring to FIG. 1, the pathological diagnostic apparatus according to the first embodiment includes an image acquiring unit 10, an image shaping unit 20, a cell nucleus shape extracting unit 30, a pathological diagnostic feature determining unit 50a, a pathological diagnostic unit 60, The diagnosis result output means 70 is provided, and the pathological diagnosis feature determination means 50a is provided with the cell nucleus feature determination means 51.

The image acquiring means 10 performs photographing on a specimen of a living tissue, converts the obtained image of the living tissue into an image file format which is a set of color data of pixels at each coordinate, and converts the image information. And memorize this.

The image shaping means 20 removes noise from the image information of the living tissue photographed and stored by the image acquiring means 10 so as to accurately determine a cell nucleus and a cavity in the image information. Performs a smoothing process and outputs smoothed image information.

The cell nucleus shape extracting means 30 individually extracts the outline of each cell nucleus from the smoothed image information output from the image shaping means 20, from this image information, and coordinates the coordinates of the pixel at the position of the outline of each cell nucleus. Is output as cell nucleus contour information.

The pathological diagnosis feature judging means 50a judges various characteristics of the living tissue used for the pathological diagnosis using various kinds of input information, and determines the characteristics such as the shape and position distribution of cell nuclei and cavities in the living tissue. Is output, and the pathological diagnosis feature determining means 50a of the first embodiment is output.
Is provided with a cell nucleus feature determining unit 51, and based on the cell nucleus contour information output from the cell nucleus shape extracting unit 30, the cell nucleus feature determining unit 51 determines the cell nucleus structural feature of the living tissue and outputs it.

The cell nucleus feature judging means 51 obtains the value of the ratio of the area of the cell nucleus area to the area of the non-cell nucleus excluding the cavity area based on the cell nucleus contour information output by the cell nucleus shape extracting means 30, and calculates the value of this ratio. The degree of cell nucleus overlap is output as a cell nucleus structural feature.

The pathological diagnosis means 60 executes pathological diagnosis based on the structural characteristic information of the living tissue output from the pathological diagnosis characteristic determination means 50. The pathological diagnosis means 60 according to the first embodiment comprises: The pathological diagnosis is executed based on the cell nucleus structural characteristics output from the pathological diagnosis characteristic determining means 50.

In the pathological diagnosis by the pathological diagnosis means 60, a pathological diagnosis is performed based on a preset judging method for cell nucleus structural features, such as judging benign when the degree of cell nucleus overlap is remarkable.

The diagnosis result output means 70 is provided for the pathological diagnosis means 6.
0 to output the pathological diagnosis result for the living tissue.

Next, the process of determining the cell nucleus structural feature by the cell nucleus feature determining means 51 of the first embodiment will be described in detail. FIG. 2 is a flowchart illustrating a process of determining a cell nucleus structure feature by the cell nucleus feature determination unit 51 according to the first embodiment.

Referring to FIG. 2, the cell nucleus structure feature determination processing by the cell nucleus feature determination means 51 of the first embodiment is performed by first setting “0” to each pixel of the image of the living tissue.
Alternatively, the array “flag [i] [j]” that takes a value of “1” is set and initialized. This obtains information on the values of the number of horizontal pixels M and the number of vertical pixels N of the image of the biological tissue, and based on the number of pixels, a binary array {flag [i] [j]: i = 1,.
M, j = 1,..., N} are set, and the values of all array elements are initialized to 0 (step 301).

Next, the position of each pixel of the image of the living tissue is
Whether it is inside or outside the area included in the cell nucleus contour (that is,
It is determined whether it is inside or outside the cell nucleus), and the array element values corresponding to all the pixels inside the cell nucleus contour are set to “1” (step 302). That is, if the pixel at the coordinates (i, j) is inside the cell nucleus contour, the corresponding array element flag [i]
Set the value of [j] to "1".

Then, each pixel (i,
j), the total number of pixels “N1”, “flag” inside the cell nucleus contour where “flag [i] [j] = 0”
[I] The total number “N2” of pixels outside the cell nucleus where [j] = 1 ”is accumulated (step 303).

The value of "N1 / N2" is determined as a cell nucleus structural feature indicating the degree of overlap of cell nuclei, and is output (step 304).

As described above, the pathological diagnosis apparatus according to the present embodiment can appropriately quantify the information on the overlap of cell nuclei and perform a pathological diagnosis based on the information on the overlap of cell nuclei.

Next, a second embodiment of the present invention will be described in detail. FIG. 3 is a block diagram illustrating a configuration of the pathological diagnosis device according to the second embodiment. In the present embodiment, a pathological diagnosis is performed based on the bias of the positional distribution of the cavity. In pathological diagnosis, uniform distribution of cavities is malignant, and uneven distribution of cavities is benign.

Referring to FIG. 3, the difference between the pathological diagnosis apparatus according to the second embodiment and the pathological diagnosis apparatus according to the first embodiment in FIG. A means 40 is provided, and a cavity feature determining means 52 is provided in place of the cell nucleus feature determining means 51 in the pathological diagnosis feature determining means 50b.

The cavity shape extracting means 40 includes the image shaping means 20
, The contour of each cavity is individually extracted from this image information, and the coordinates of the pixel at the position of the cavity contour of each cavity are output as cavity contour information.

The pathological diagnosis feature judging means 50b of the second embodiment includes a cavity feature judging means 52, and the cavity structure judging means 52 uses the cavity feature judging means 52 based on the cavity contour information outputted by the cavity shape extracting means 40. Judge the feature and output it.

The cavity feature determining means 52 includes a cavity shape extracting means 4
0, based on the cavity contour information output, performs enlarging processing and morphology processing of the cavity contour in the region of interest, thereby determining the deviation of the position distribution of the cavity contour in the region of interest and determining the position distribution of the cavity contour. Is output as the cavity structure feature.

The pathological diagnosis means 60 according to the second embodiment comprises:
The pathological diagnosis is executed based on the cavity structure characteristics output from the pathological diagnostic feature determination means 50b.

In the pathological diagnosis by the pathological diagnosis means 60, a pathological diagnosis is performed based on a predetermined judging method for the cavity structure feature, such as judging malignant when the cavities are evenly distributed in the region.

Next, the process of determining the cavity structure feature by the cavity feature determination means 52 of the second embodiment will be described in detail. FIG. 4 is a flowchart for explaining the determination process of the cavity structure feature by the cavity feature determination unit 52 according to the second embodiment.

Referring to FIG. 4, first, the number N of cavities in the region of the living tissue to be diagnosed is determined, and a distance threshold T relating to the contour enlarging process described below is set (step 5).
01).

An array ｛(Lx
[I] [j], Ly [i] [j]): i = 1,..., N,
j = 1,...}, and the coordinates (x, y) of the contour of the closed region including the whole cavity contour stored here are arranged in an array {(Dx
[K], Dy [k]): k = 1,... (Step 502).

The contour list L for enlarging all the cavity contours
(Step 503).

All the cavity contours in contact with the closed region contour in the enlargement processing contour list L are deleted (step 504).

All the contours of the cavity whose distance from the contour of another adjacent cavity is equal to or smaller than the threshold value T set in step 501 are listed in a list L.
(Step 505).

It is checked whether or not the cavity contour remains in the list L (step 506). If the contour remains, the cavity shape is enlarged by one pixel toward the outside of the cavity in the list L toward the outside of the cavity. (Step 507), the process proceeds to Step 5.
Return to 04.

The processing for enlarging the cavity shape in steps 504 to 507 is executed for all the cavity contours in the list L.

The total number of pixels S in the closed area and the total number of pixels S 'in the enlarged contour are accumulated (step 508).

It is determined that S ′ / (S−S ′) is a cavity structure feature representing the cavity distribution in the closed area, and this is output (step 5).
09).

Next, with reference to FIGS. 5 to 8, the process of enlarging the cavity shape by the cavity feature determining means 52 of the present embodiment will be described with reference to a specific example of a biological tissue image.

FIG. 5 is an example of an image of a biological tissue sample showing a malignant cavity distribution, FIG. 6 is an example of an image of a biological tissue sample showing a benign cavity distribution, and FIG. FIG. 8 is an image of the biological tissue sample image showing the distribution after the cavity shape enlarging process, and FIG. 8 is an image of the biological tissue sample image showing the benign cavity distribution of FIG. 6 after the cavity shape enlarging process.

When FIG. 6 and FIG. 7 are compared, in FIG. 6, white regions uniformly representing the cavity structure are distributed, indicating malignancy.

The white area, which is the cavity structure of these two images, is subjected to cavity shape enlargement processing by the cavity feature determination means 52 of the present embodiment, and FIGS. 8 and 9 are obtained.

When FIG. 8 and FIG. 9 are compared, in FIG. 8 in which the cavities are uniformly distributed in FIG. 6 before the processing, the white region having the cavity structure as a whole is enlarged in the image. In FIG. 9 in which the distribution of the cavity is biased in FIG. 7 before the processing, the white region as the cavity structure is enlarged only in a part of the image.

As a result, the difference in the uniformity of the distribution of the white area as the cavity structure in FIGS. 6 and 7 before the processing can be expressed as the difference in the area of the white area by the enlargement processing.

As described above, in the present embodiment, information on the uniformity of the distribution area of the cavity, which is difficult to determine except by a person, is represented by the difference in the area of the white region by the enlargement processing of the cavity structure. This allows the diagnostic apparatus to automatically determine the uniformity of the cavity distribution state by comparing the areas of the white region and the black region.

In addition, the information on the uniformity of the distribution state of the cavity can be appropriately quantified, and a pathological diagnosis can be made based on the information on the uniformity of the distribution state of the cavity.

Next, a third embodiment of the present invention will be described in detail. FIG. 9 is a block diagram illustrating a configuration of a pathological diagnosis device according to the third embodiment.

In the present embodiment, a pathological diagnosis is performed based on the cell nucleus-cavity mutual structure feature representing the distance between the cavity and the cell nuclei around the cavity based on the individual cell nucleus contour information and the cavity contour information. According to pathological diagnosis, in a malignant living tissue, there is a certain large distance between a cavity and a cell nucleus around the cavity.

Referring to FIG. 9, the pathological diagnosis apparatus according to the third embodiment is different from the pathological diagnosis apparatus according to the first embodiment in FIG. 1 and the pathological diagnosis apparatus according to the second embodiment in FIG. The difference is that the cell nucleus shape extraction means 30
In addition, a cavity shape extraction means 40 is provided, and a cell nucleus-cavity mutual feature determination means 53 is provided in the pathological diagnosis feature determination means 50c.

The cell nucleus-cavity mutual feature judging means 53 uses the cell nucleus contour information output from the cell nucleus shape extracting means 30 and the cavity contour information means output from the cavity shape extracting means 40.
The nucleus-cavity mutual structure feature is determined from the circumscribed rectangle of the cavity contour and the circumscribed rectangle of the nearest cell nucleus group of the cavity contour.

FIG. 10 is a flowchart for explaining the process of determining the mutual feature of the cell nucleus and the cavity by the cell nucleus and cavity mutual feature determining means of the third embodiment.

First, one cavity is selected (step 110).
1).

The circumference ln of the circumscribed rectangle of the contour of the selected cavity is obtained (step 1102).

Of the cell nuclei around the contour of the selected cavity,
A cell nucleus close to the cavity contour is selected as the nearest cell nucleus group, and the number CN is counted, and the perimeter of the circumscribed rectangle of the contour of each cell nucleus included in the nearest cell nucleus array is arrayed.
[I]: Stored in i = 1,..., CN} (Step 11)
03).

The average value CL of the circumscribed rectangle circumference of the cell nucleus contour of the nearest cell nucleus group is obtained by obtaining the average value of the elements of the array {C [i]: i = 1,..., CN} (step 1104). .

The perimeter LN of the circumscribed rectangular area including the nearest cell nucleus group is determined (step 1105).

(LN-ln) / CL is determined as a cell nucleus-cavity mutual structural feature representing the distance between the cavity and the surrounding cell nuclei (step 1106).

FIG. 11 is a diagram showing an example of the determination of the perimeter of the circumscribed rectangular area of the cavity contour and the perimeter of the circumscribed rectangular area of the nearest cell nucleus group by the cell nucleus-cavity mutual feature determination means 53 of this embodiment. .

The cell nucleus-cavity mutual feature judging means 53 is shown in FIG.
As shown in FIG. 1, the circumference ln of the circumscribed rectangular area of the cavity contour
Then, the circumference LN of the circumscribed rectangular region of the cell nucleus group closest to the cavity is measured.

As described above, the perimeter ln of the circumscribed rectangular area of the cavity contour and the perimeter LN of the circumscribed rectangular area of the cell nucleus group closest to the cavity are obtained.
And the difference (LN-ln) is normalized by the average size CL of the nearest cell nucleus group, so that the cell nucleus representing the distance between the cavity and the surrounding cell nuclei regardless of the magnification of the image.
The cavity interaction feature (LN-ln) / CL can be determined.

As described above, the pathological diagnosis apparatus of the present embodiment appropriately quantifies the information on the positional relationship between the cavity and the cell nuclei around the cavity using the circumscribed rectangular region circumference of the cavity and the cell nucleus group. Pathological diagnosis based on information on the positional relationship between the cavity and the surrounding cell nuclei.

Next, a fourth embodiment of the present invention will be described in detail. FIG. 12 is a block diagram illustrating a configuration of a pathological diagnosis apparatus according to the fourth embodiment.

As in the third embodiment, the present embodiment is based on the cell nucleus-cavity mutual structure feature representing the distance between the cavity and the surrounding cell nuclei based on the individual cell nucleus contour information and the cavity contour information. A pathological diagnosis is made.

Referring to FIG. 12, the difference between the pathological diagnosis apparatus according to the fourth embodiment and the pathological diagnosis apparatus according to the third embodiment in FIG.
0d is provided with a cell nucleus-cavity mutual feature determination means 53a.

The cell nucleus-cavity mutual feature judging means 53 a uses the cell nucleus contour information output from the cell nucleus shape extracting means 30 and the cavity contour information means output from the cavity shape extracting means 40 to obtain the third embodiment. Cell nucleus-cavity mutual feature determination means 5
Unlike the case of No. 3, the cell nucleus-cavity mutual structure characteristic representing the positional relationship between the cavity and the cell nucleus is determined by calculating the area.

FIG. 13 is a flowchart for explaining the process of determining the mutual structural feature between the cell nucleus and the cavity by the cell nucleus-cavity mutual feature determining means 53a of the fourth embodiment.

First, one cavity is selected (step 140).
1).

The area S of the contour of the selected cavity is obtained (step 1402).

Of the cell nuclei around the contour of the selected cavity,
The cell nuclei in close proximity to the cavity contour are selected as the nearest cell nucleus group, and the number CN is counted, and the area of the circumscribed rectangle of the contour of each cell nucleus included in the nearest cell nucleus array is arrayed.
[I]: Stored in i = 1,..., CN # (Step 14)
03).

The average value CL of the circumscribed rectangular area of the cell nucleus contour of the nearest cell nucleus group is obtained by obtaining the average value of the elements of the array {C [i]: i = 1,..., CN} (step 1404).

The area S 'of the closed region formed by connecting the centers of the adjacent cell nuclei of the group of closest cell nuclei is obtained (step 14).
05).

(S'-S) / CL is determined as a cell nucleus-cavity mutual structural feature representing the distance between the cavity and the cell nuclei around the cavity (step 1406).

As described above, the pathological diagnosis apparatus according to the present embodiment appropriately quantifies the information on the positional relationship between the cavity and the cell nuclei around the cavity by using the area of the cavity, the cell nucleus, and the like. A pathological diagnosis can be performed based on information on the positional relationship between the cell and its surrounding cell nuclei.

The first embodiment of the present invention has been described with an example in which the cell nucleus feature determining means 51 determines the degree of cell nucleus overlap based on the area ratio between the cell nucleus region and the non-cell nucleus region.
The degree of overlap of the cell nuclei can also be determined using the distribution of the distance between the centers of gravity of the adjacent cell nuclei.

That is, instead of calculating the area ratio, the average value of the distance between the centers of gravity of the cell nuclei and the like is determined as the degree of overlap of the cell nuclei. The more the cell nuclei overlap, the more this is judged to be benign.

According to the second embodiment of the present invention, in the determination of the cavity structure feature by the cavity feature determination unit 52, S ′ / (S −S ′) defines the deviation of the cavity position distribution, but may be defined as S ′ / S.

In each case, the lower the bias value of the cavity position distribution, the more the cavity position distribution is biased. However,
In the case of defining by S ′ / S, the value of the bias of the cavity position distribution can always be represented by a positive number of 1 or less.

The second embodiment of the present invention has been described with an example in which the cavity feature determination means 52 determines the position distribution of the cavity contour by using the cavity contour enlarging process. Alternatively, the determination can be made using the distance between the centers of gravity.

Regarding the third embodiment and the fourth embodiment of the present invention, the cell nucleus-cavity mutual feature judging means 53, 53
a described the mutual structural feature between the cell nucleus and the cavity using the circumference and area of the circumscribed rectangle of the cavity contour and the circumscribed rectangle of the nearest cell nucleus group. The determination can be made using the distance.

In the above-described pathological diagnosis feature determining means 50, the cell nucleus feature determining means 51, the cavity feature determining means 52, and the cell nucleus-cavity mutual feature determining means 53 have only one kind of various methods for determining image characteristics. Instead, a plurality of features determination methods as described above may be combined.

Further, in the pathological diagnosis characteristic determining means 50, the cell nucleus characteristic determining means 51, the cavity characteristic determining means 52, and the cell nucleus-cavity mutual characteristic determining means 53 are not limited to those having only one type, but may be any one. The above determination means may be provided. As an example of this, FIG.
Within 0, the cell nucleus feature determining means 51 and the cavity feature determining means 52
1 is a block diagram showing a configuration of a pathological diagnosis apparatus according to an embodiment of the present invention, which includes a cell nucleus-cavity mutual feature determination unit 53. FIG.

In the above embodiment, the image shaping means 20, the cell nucleus shape extracting means 30, and the cavity shape extracting means 40
Performs preprocessing for pathological diagnosis feature determination processing as a part independent of the pathological diagnosis feature determination means 50. These means include a pathological diagnosis feature determination means 50, a cell nucleus feature determination means 51, and a cavity feature The determination unit 52 and the cell nucleus-cavity mutual feature determination unit 53 may be provided.

In the above embodiment, the pathological diagnosis means 60 outputs a diagnosis result based on the various characteristics determined by the pathological diagnosis characteristic determination means 50. The values of the various characteristics are combined with the diagnosis results. May be output. Further, the values of the various characteristics determined by the pathological diagnosis characteristic determining unit 50 may be output without the pathological diagnosis unit 60, and the diagnosis may be performed by a doctor or the like based on the output various characteristics.

In the above embodiment, the image acquisition means 10 performs photographing on a specimen of a living tissue, and converts this information into an image file composed of electrical signals. Or the like may be used as a diagnosis target. In this case, processing after photographing of a biological tissue sample or the like by the image acquisition unit 10 is performed.

Similarly, the image may be converted into an image file composed of electrical signals in advance, and the image stored in a recording medium such as a hard disk or a floppy disk may be used as a diagnosis target. In this case, it is not necessary to provide the image acquisition unit 10.

Although the present invention has been described with reference to the preferred embodiments and examples, the present invention is not necessarily limited to the above embodiments and examples, and various modifications may be made within the scope of the technical idea. Modifications can be made.

[0114]

According to the pathological diagnosis apparatus of the present invention as described above, the following effects can be achieved.

First, by appropriately quantifying information such as the positional relationship and distribution state of cell nuclei and cavities in a biological tissue image, it is possible to use such information for pathological diagnosis by a pathological diagnosis apparatus. Therefore, it is possible to perform a pathological diagnosis of a type that has conventionally required the judgment of a doctor or the like and cannot be performed by a pathological diagnostic apparatus, or a pathological diagnosis with high accuracy.

Second, the information on the overlapping of cell nuclei is appropriately quantified, and a pathological diagnosis can be made based on the information on the overlapping of cell nuclei.

Third, information on the distribution state of the cavity is appropriately quantified, and a pathological diagnosis can be made based on the information on the distribution state of the cavity.

Fourth, information on the positional relationship between the cavity and the cell nucleus located closest to the cavity is appropriately quantified, and a pathological diagnosis can be made based on the information on the positional relationship between the cavity and the cell nucleus.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a pathological diagnosis apparatus according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating a process of determining a cell nucleus structural feature by a cell nucleus feature determining unit according to the first embodiment of the present invention.

FIG. 3 is a block diagram illustrating a configuration of a pathological diagnosis device according to a second embodiment of the present invention.

FIG. 4 is a flowchart illustrating a cavity structure feature determination process performed by a cavity feature determination unit according to the second embodiment of this invention.

FIG. 5 is an example of an image of a biological tissue specimen showing a malignant cavity distribution.

FIG. 6 is an example of an image of a biological tissue specimen showing a benign cavity distribution.

FIG. 7 is an image of the biological tissue specimen image showing the malignant cavity distribution shown in FIG. 5 after the cavity shape enlarging process by the cell nucleus feature determination unit according to the second embodiment of the present invention.

FIG. 8 is an image of the biological tissue specimen image showing the benign cavity distribution shown in FIG. 6 after the cavity shape enlarging process performed by the cell nucleus feature determination unit according to the second embodiment of the present invention.

FIG. 9 is a block diagram illustrating a configuration of a pathological diagnosis apparatus according to a third embodiment of the present invention.

FIG. 10 is a flowchart illustrating a process of determining a cell nucleus-cavity mutual structure feature by a cell nucleus-cavity mutual feature determination unit according to the third embodiment of the present invention.

FIG. 11 is a diagram illustrating an example of determination of a perimeter of a circumscribed rectangular region of a cavity contour and a perimeter of a circumscribed rectangular region of a nearest cell nucleus group by the cell nucleus-cavity mutual feature determination unit according to the third embodiment of this invention; is there.

FIG. 12 is a block diagram illustrating a configuration of a pathological diagnosis apparatus according to a fourth embodiment of the present invention.

FIG. 13 is a flowchart illustrating a process of determining a cell nucleus-cavity mutual structure feature by a cell nucleus-cavity mutual feature determination unit according to the fourth embodiment of the present invention.

FIG. 14 is a block diagram showing a configuration of a pathological diagnosis apparatus according to another embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Image acquisition means 20 Image shaping means 30 Cell nucleus shape extraction means 40 Cavity shape extraction means 50 Pathology diagnosis feature determination means 51 Cell nucleus feature determination means 52 Cavity feature determination means 53 Cell nucleus-cavity mutual feature determination means 60 Pathology diagnosis means 70 Diagnosis result output means

 ────────────────────────────────────────────────── ─── Continued on the front page F term (reference) 2G045 AA24 AA25 CB01 FA19 GB02 GB04 GB10 JA01 JA07 5B057 AA10 DA03 DA07 DB02 DC02 DC03 DC04 DC06

Claims (12)

[Claims] 1. A pathological diagnosis feature determining unit that determines a feature used for pathological diagnosis from an image of a living tissue, wherein the pathological diagnostic feature determining unit determines a characteristic based on a positional relationship or a distribution state of a cell nucleus or a cavity in the image. A pathological diagnosis apparatus characterized in that is determined by a numerical value. 2. The pathological diagnosis apparatus according to claim 1, wherein said pathological diagnosis feature determining means includes a cell nucleus characteristic determining means for determining, as a cell nucleus structural feature, the degree of overlap of the entire cell nucleus. 3. The pathological diagnosis apparatus according to claim 2, wherein said cell nucleus feature determining means obtains an area occupied by all cell nuclei in the image, and determines the degree of overlap using the area. 4. The method according to claim 1, wherein said cell nucleus feature determining means obtains an interval between centroids of said cell nuclei of adjacent cell nuclei in said image, and judges said degree of overlap using said interval. 3. The pathological diagnosis apparatus according to 2. 5. The pathological diagnosis apparatus according to claim 1, wherein said pathological diagnosis characteristic determining means includes a cavity characteristic determining means for determining a distribution state of the cavity as a cavity structure characteristic. 6. The pre-enlargement image is obtained by enlarging a portion indicating a cavity in the image with respect to the image, and determining an area occupied by the portion indicating the cavity in the enlarged image. Performing a cavity shape enlarging process expressing information on the distribution state of the cavity in the cavity, obtaining an area occupied by a portion indicating the cavity in the image after the cavity shape enlarging process, and determining the distribution state using the area. The pathological diagnosis apparatus according to claim 5, wherein: 7. The cavity shape enlarging process performed by the cavity feature judging means may be such that the image does not touch a boundary of an enlarging process target range and an interval with another cavity is equal to or greater than a predetermined threshold value. Performing a selection process of selecting all the cavities in the image as cavities to be subjected to enlargement processing. Next, for each of the cavities to be subjected to enlargement processing, a region within a range defined outward from the outline of the cavity is selected. An image rewriting process for newly including pixels inside the cavity is performed, and the above-described image is subjected to the above-mentioned selection process again, until the cavity selected by the above-mentioned selection process disappears.
7. The pathological diagnosis apparatus according to claim 6, wherein the cavity shape is enlarged by repeating a series of processes of the selection process and the image rewriting process. 8. The apparatus according to claim 1, wherein said cavity characteristic determining means determines a distance between the centers of gravity of the individual chambers in the image, and determines the distribution state using the distance between the centers of gravity. 6. The pathological diagnosis apparatus according to 5. 9. A cell nucleus for determining a positional relationship between the cavity and a group of closest cell nuclei constituted by a plurality of cell nuclei existing around the cavity as a cell nucleus-cavity mutual structure characteristic. The pathological diagnosis apparatus according to claim 1, further comprising: a cavity mutual characteristic determination unit. 10. The cell nucleus-cavity mutual feature determination means obtains a length of a circumscribed rectangle of the outline of the cavity and a length of a circumscribed rectangle of the group of nearest cell nuclei, and determines the length of the rectangle circumscribed of the cavity contour and the nearest cell nucleus. The pathological diagnosis apparatus according to claim 9, wherein the positional relationship between the cavity and the nearest cell nucleus group is determined using a group circumscribed rectangle length. 11. The cell nucleus-cavity mutual feature determination means, comprising: an area of the cavity, an area of an individual cell nucleus of the group of the closest cell nuclei, and a closed region connecting a center of gravity of each cell nucleus of the group of the closest cell nucleus. The area of the cavity, the area of the cell nucleus, and the area of the closed region are determined, and the positional relationship between the cavity and the nearest cell nucleus group is determined using the area of the cavity. Pathological diagnostic device. 12. The cell nucleus-cavity mutual feature determination means obtains a distance between a contour of the cavity and individual cells of the nearest cell nucleus group, and uses the distance to determine the distance between the cavity and the nearest cell nucleus. The pathological diagnosis apparatus according to claim 9, wherein the positional relationship with a group is determined.