CN106327508B - Ki67 index automatic analysis method - Google Patents
Ki67 index automatic analysis method Download PDFInfo
- Publication number
- CN106327508B CN106327508B CN201610710869.4A CN201610710869A CN106327508B CN 106327508 B CN106327508 B CN 106327508B CN 201610710869 A CN201610710869 A CN 201610710869A CN 106327508 B CN106327508 B CN 106327508B
- Authority
- CN
- China
- Prior art keywords
- region
- cell
- positive
- index
- expression
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20152—Watershed segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a kind of Ki67 index automatic analysis methods comprising following steps: S10, image preprocessing；S20, hot spot region screening；S30, cell number calculate；S40, result output.Ki67 index automatic analysis method of the invention has preferable reproducibility, the digital picture of Ki67 is analyzed by computer automatic batch, the acquisition Ki67 index of stability and high efficiency, and negative and positive nucleus is marked by different colours in digital picture, hot spot region is quickly identified by algorithm and to the yin and yang attribute nuclei count in hot spot region, medical worker is helped more accurately to analyze histopathology correlated characteristic.
Description
Technical field
The present invention relates to the statistical method of data more particularly to a kind of Ki67 index automatic analysis methods.
Background technique
It is related to the investigation of mitotic index in the diagnosis of neuroendocrine tumor, and Ki67 index can be in certain journey
Reflect the mitosis situation in tissue on degree, in the Ki67 Index Assessment of WHO publication, doctor needs to investigate tissue middle-jiao yang, function of the spleen and stomach
Highest region is expressed, evaluates the Ki67 positive events of tissue by counting 500~2000 tumour cells.In real process
In, doctor quickly examines out the hot spot region in slice under low power lens, then by 10 high power field of view in hot spot region
Under yin and yang attribute cell count assess Ki67 index.
At present to two kinds of means are mainly used in terms of Ki67 Index Assessment, it is respectively: 1) exports hot spot region for picture,
It prints doctor and counts 500~2000 tumour cells in hot spot region to obtain Ki67 cell and negative cells, pass through
Positive nucleus number percent describes Ki67 index；2) doctor passes through the high power lens view of ascites hot spot region 10
Yezhong yin and yang attribute nucleus amount is to evaluate Ki67 index.
By the comparison to above two method, above-mentioned first way time and effort consuming but result most can really react
Ki67 index；The second way is the most frequently used but result reproducibility is poor, and different physicians may provide inconsistent result and needs
Doctor accumulates relevant clinical experience.
And there is no the yin and yang attribute cell nucleus numbers in a kind of hot spot region selected by computer automatic analysis at present
Amount, thus obtain positive nucleus number percent, the method for completing the evaluation to slice.
Summary of the invention
It is an object of the present invention to provide a kind of Ki67 index automatic analysis methods, analyze Ki67 by computer automatic batch
Digital picture, the acquisition Ki67 index of stability and high efficiency, and negative by different colours label in digital picture and the positive is thin
Karyon helps medical worker more accurately to analyze histopathology correlated characteristic.
The present invention solves technical problem and adopts the following technical scheme that a kind of Ki67 index automatic analysis method comprising with
Lower step:
S10, the digital picture that original magnification is 40 times is narrowed down to 5 times；
S20, with 0.2mm2Region is that highest 10 regions of Ki67 index are chosen in sampling region, with 2mm2Region is sampling
Choose highest 1 region of Ki67 index in region；
S30, to selected 10 0.2mm2Region and 1 2mm2Region identifies the Ki67 positive under original magnification
The cell of the cell of expression and negative expression, and the cell of the cell of Ki67 positive expression and negative expression is counted；
S40, obtain corresponding region Ki67 positive expression cell number percentage when corresponding area percentage.
Optionally, the step S20 includes:
S201, be side length according to certain pixel quantity by Digital Image Segmentation it is several small cubes, and makes described small
The area of square is 0.2mm2；Be side length according to certain pixel quantity by Digital Image Segmentation it is several small cubes, and makes
The area of the small cube is 2mm2；
S202, using each vertex of each small cube as the center of circle, calculate certain radius within the scope of positive region exist
Shared area ratio in the border circular areas；
S203, the positive region area hundred for being more entirely sliced border circular areas representated by each vertex of all small cubes
Divide ratio, from 0.2mm2Region in select 10 positive region area percentages highest as hot spot region, and from 2mm2's
It is highest as hot spot region that 1 positive region area percentage is selected in region.
Optionally, the step S30 specifically:
S301, edge enhancing and optimization are carried out to the edge of selected hot spot region；
S302, by watershed algorithm, to step S301, treated that digital picture is split, and obtains all nucleus；
S303, cell and the negative Cellular spectroscopic information gap expressed according to Ki67 positive expression, by yin and yang attribute cell
It distinguishes in different colors, realizes the mark of the cell to the cell and negative expression of Ki67 positive expression, and to the Ki67 positive
The cell of expression and the cell of negative expression are counted.
The invention has the following beneficial effects: Ki67 index automatic analysis methods of the invention to have preferable reproducibility,
The digital picture of Ki67 is analyzed by computer automatic batch, the acquisition Ki67 index of stability and high efficiency simultaneously passes through in digital picture
Different colours label is negative and positive nucleus, and hot spot region and thin to the yin and yang attribute in hot spot region is quickly identified by algorithm
Karyon counts, and medical worker is helped more accurately to analyze histopathology correlated characteristic.
Detailed description of the invention
Fig. 1 is the digital picture (pixel: 27053 × 19259 that original magnification of the invention is 40 times；Resolution ratio:
0.23μm/pixel)；
Fig. 2 is the digital picture (pixel: 3382 × 2408) that amplification factor of the invention is 5 times；
Fig. 3 is that the present invention passes through 0.2mm2Region segmentation after digital picture；
Fig. 4 is that the present invention passes through 2mm2Region segmentation after digital picture；
Fig. 5 (a) is original image of the invention；It is the result images (yellow after the screening of hot spot region on the right side of Fig. 5 (b)
The part of mark is positive cell range)；
Fig. 6 is hot spot region selection result schematic diagram (0.2mm of the invention2Region)；
Fig. 7 is hot spot region selection result schematic diagram (2mm of the invention2Region).
Fig. 8 (a) is the original image after the selection of hot spot region；Fig. 8 (b) is the image after mean filter；
Specific embodiment
Technical solution of the present invention is further elaborated below with reference to examples and drawings.
Embodiment 1
A kind of Ki67 index automatic analysis method is present embodiments provided, the hot spot region in identification digital picture is passed through
It realizes the automated analysis of computer Ki67 index, helps the doctor preferably to evaluate Ki67 positive degree；And including
Following step:
S10, image preprocessing.
Analysis in the present embodiment, for Ki67 index, it is necessary first to realize the identification of hot spot region in digital image；
For this reason, it may be necessary to pre-processed to digital picture, still, when the amplification factor of digital picture is 40, due to its data volume compared with
Greatly, cause the time-consuming when hot spot region is screened more, carried out therefore, it is necessary to the digital picture for being 40 times by original magnification etc.
It is scaled to suitable amplification factor, then carries out hot spot region screening again, in the present embodiment, by largely testing table
It is bright, be suitably when the digital picture that original magnification is 40 times is narrowed down to 5 times, i.e., in the present embodiment, described image is pre-
Processing step specifically: the digital picture that original magnification is 40 times is narrowed down to 5 times.
S20, hot spot region screening.
After pre-processing the digital picture, need to screen the hot spot region in digital picture, but
It is, since the highest region of Ki67 index can be able to satisfy 10 high power field of view (about 2mm in slice2), it is also possible to inadequate 2mm2,
It needs additionally to be chosen, it therefore, can be with 0.2mm in the present embodiment2Region is that selection Ki67 index in sampling region is highest
10 regions, with 2mm2Region is that highest 1 region of Ki67 index is chosen in sampling region.
In the present embodiment, the specific steps of the hot spot region screening are as follows:
S201, be side length according to certain pixel quantity by Digital Image Segmentation it is several small cubes, and makes described small
The area of square is 0.2mm2；
S202, using each vertex of each small cube as the center of circle, calculate certain radius within the scope of positive region exist
Shared area ratio in the border circular areas (region within the scope of certain radius)；
S203, the positive region area hundred for being more entirely sliced border circular areas representated by each vertex of all small cubes
Divide ratio, it is highest as hot spot region from 10 positive region area percentages are wherein selected.
It is highest as 2mm also to can choose 1 positive region area percentage for similarly mode2Hot spot region.
S30, cell number calculate.
After step S20 execution, that is, start to calculate cell number, the cell number includes Ki67 positive expression
The cell number of cell number and negative expression can be to selected 10 0.2mm in the present embodiment2Region and 1 2mm2Region
The cell of Ki67 positive expression and the cell of negative expression are identified under original magnification (40 times), and to Ki67 positive expression
Cell and the cell of negative expression counted.
In the present embodiment, the specific steps of the cell number calculating are as follows:
S301, edge enhancing and optimization are carried out by edge of the mean bias filtering to selected hot spot region；
S302, by watershed algorithm, to step S301, treated that digital picture is split, and obtains all nucleus；
Yin and yang attribute cell is filtered out by the mean value of the channel R (Layer 1) to be identified, using apart from map to yin and yang attribute nucleus
Carry out experiment material.
S303, cell and the negative Cellular spectroscopic information gap expressed according to Ki67 positive expression, by yin and yang attribute cell
It distinguishes in different colors, realizes the mark of the cell to the cell and negative expression of Ki67 positive expression, and to the Ki67 positive
The cell of expression and the cell of negative expression are counted.
It is highly preferred that the step S301 is specifically included:
S40, result output, result is as shown in table one and table two:
One: 10 0.2mm of table2Cell image processing result figure
Table two: 2mm2Cell image processing result figure
After completing step S30, it can the cell number percentage for obtaining the Ki67 positive expression of corresponding region is when right
The area percentage answered.The Ki67 positive cell percentage can be by the total number of cells of Ki67 positive expression divided by Ki67 sun
Property total number of cells of expression and the sum of the total number of cells of negative expression；The area percentage of the Ki67 positive expression can pass through
The pixel of Ki67 positive expression cell is divided by the pixel of the cell of Ki67 positive expression and the sum of the pixel of cell of negative expression.
Ki67 index automatic analysis method of the invention has preferable reproducibility, is analyzed by computer automatic batch
The digital picture of Ki67, the acquisition Ki67 index of stability and high efficiency simultaneously mark the negative and positive by different colours in digital picture
Nucleus is quickly identified hot spot region by algorithm and to the yin and yang attribute nuclei count in hot spot region, helps medication work
Person more accurately analyzes histopathology correlated characteristic.
The sequencing of above embodiments is not only for ease of description, represent the advantages or disadvantages of the embodiments.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (3)
1. a kind of Ki67 index automatic analysis method, which comprises the following steps:
S10, the digital picture that original magnification is 40 times is narrowed down to 5 times；
S20, with 0.2mm2Region is that highest 10 regions of Ki67 index are chosen in sampling region, with 2mm2Region is sampling region
Choose highest 1 region of Ki67 index；
S30, to selected 10 0.2mm2Region and 1 2mm2Region identifies Ki67 positive expression under original magnification
Cell and negative expression cell, and the cell of the cell of Ki67 positive expression and negative expression is counted；
S40, obtain corresponding region Ki67 positive expression cell number percentage when corresponding area percentage.
2. Ki67 index automatic analysis method according to claim 1, which is characterized in that the step S20 includes:
S201, be side length according to certain pixel quantity by Digital Image Segmentation it is several small cubes, and makes the small cube
Area be 0.2mm2；Be side length according to certain pixel quantity by Digital Image Segmentation it is several small cubes, and makes described
The area of small cube is 2mm2；
S202, using each vertex of each small cube as the center of circle, calculate certain radius within the scope of positive region in circle
Shared area ratio in region；
S203, the positive region area percentage for being more entirely sliced border circular areas representated by each vertex of all small cubes
Than from 0.2mm2Region in select 10 positive region area percentages highest as hot spot region, and from 2mm2Area
It is highest as hot spot region that 1 positive region area percentage is selected in domain.
3. Ki67 index automatic analysis method according to claim 1 or 2, which is characterized in that the step S30 is specific
Are as follows:
S301, edge enhancing and optimization are carried out to the edge of selected hot spot region；
S302, by watershed algorithm, to step S301, treated that digital picture is split, and obtains all nucleus；
S303, cell and the negative Cellular spectroscopic information gap expressed according to Ki67 positive expression, by yin and yang attribute cell with not
It is distinguished with color, realizes the mark of the cell to the cell and negative expression of Ki67 positive expression, and to Ki67 positive expression
Cell and the cell of negative expression counted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610710869.4A CN106327508B (en) | 2016-08-23 | 2016-08-23 | Ki67 index automatic analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610710869.4A CN106327508B (en) | 2016-08-23 | 2016-08-23 | Ki67 index automatic analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106327508A CN106327508A (en) | 2017-01-11 |
| CN106327508B true CN106327508B (en) | 2019-04-02 |
Family
ID=57742155
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610710869.4A Expired - Fee Related CN106327508B (en) | 2016-08-23 | 2016-08-23 | Ki67 index automatic analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106327508B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106846310A (en) * | 2017-01-19 | 2017-06-13 | 宁波江丰生物信息技术有限公司 | A kind of pathology aided analysis method based on immunohistochemistry technique |
| CN112215790A (en) * | 2019-06-24 | 2021-01-12 | 杭州迪英加科技有限公司 | KI67 index analysis method based on deep learning |
| CN110780080B (en) * | 2019-11-08 | 2024-07-12 | 安邦（厦门）生物科技有限公司 | Blood type analyzer and blood type analysis method |
| CN111413504B (en) * | 2020-04-03 | 2022-01-28 | 河北医科大学第四医院 | Standard comparison card for assisting interpretation of KI67 proliferation index |
| CN114299490B (en) * | 2021-12-01 | 2024-03-29 | 万达信息股份有限公司 | Tumor microenvironment heterogeneity evaluation method |
| CN114494204A (en) * | 2022-01-27 | 2022-05-13 | 复旦大学 | Ki67 index calculation method based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102254150A (en) * | 2005-10-14 | 2011-11-23 | 尤尼森斯繁殖技术公司 | Determination of a change in a cell population |
| CN104122189A (en) * | 2014-08-07 | 2014-10-29 | 云南中烟工业有限责任公司 | In-vitro cell micronuclei method for detecting total particulate matters of main stream smoke of cigarettes |
| CN104813366A (en) * | 2012-12-04 | 2015-07-29 | 通用电气公司 | Systems and methods for using immunostaining mask to selectively refine ISH analysis results |
| CN105719294A (en) * | 2016-01-21 | 2016-06-29 | 中南大学 | Breast cancer pathology image mitosis nucleus automatic segmentation method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| ES2573945T3 (en) * | 2008-09-16 | 2016-06-13 | Novartis Ag | Reproducible quantification of biomarker expression |
-
2016
- 2016-08-23 CN CN201610710869.4A patent/CN106327508B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102254150A (en) * | 2005-10-14 | 2011-11-23 | 尤尼森斯繁殖技术公司 | Determination of a change in a cell population |
| CN104813366A (en) * | 2012-12-04 | 2015-07-29 | 通用电气公司 | Systems and methods for using immunostaining mask to selectively refine ISH analysis results |
| CN104122189A (en) * | 2014-08-07 | 2014-10-29 | 云南中烟工业有限责任公司 | In-vitro cell micronuclei method for detecting total particulate matters of main stream smoke of cigarettes |
| CN105719294A (en) * | 2016-01-21 | 2016-06-29 | 中南大学 | Breast cancer pathology image mitosis nucleus automatic segmentation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106327508A (en) | 2017-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106327508B (en) | Ki67 index automatic analysis method | |
| US10565479B1 (en) | Identifying and excluding blurred areas of images of stained tissue to improve cancer scoring | |
| Merino et al. | Optimizing morphology through blood cell image analysis | |
| US8486704B2 (en) | Methods of chromogen separation-based image analysis | |
| US9239281B2 (en) | Method and device for dividing area of image of particle in urine | |
| CN112215790A (en) | KI67 index analysis method based on deep learning | |
| CN109074645B (en) | Spatial index creation for IHC image analysis | |
| Nurzynska et al. | CAS: Cell annotation software–research on neuronal tissue has never been so transparent | |
| US20210012088A1 (en) | Method for detection of cells in a cytological sample having at least one anomaly | |
| US20230260125A1 (en) | Digital pathology artificial intelligence quality check | |
| Lezoray et al. | Segmentation of cytological images using color and mathematical morphology | |
| CN113724235B (en) | Semi-automatic Ki67/ER/PR negative and positive cell counting system and method under condition of changing environment under mirror | |
| Irshad et al. | Spectral band selection for mitosis detection in histopathology | |
| WO2014181024A1 (en) | Computer-implemented method for recognising and classifying abnormal blood cells, and computer programs for performing the method | |
| Liu et al. | Faster r-cnn based robust circulating tumor cells detection with improved sensitivity | |
| Aubreville et al. | Augmented mitotic cell count using field of interest proposal | |
| CN112651368B (en) | DNA ploidy analysis method for large-size cell microscopic image | |
| Amitha et al. | Developement of computer aided system for detection and classification of mitosis using SVM | |
| Aubreville et al. | Field of interest proposal for augmented mitotic cell count: comparison of two convolutional networks | |
| CN113408527B (en) | Efficient PM (particulate matter) based on image fusion characteristics 2.5 Concentration prediction method | |
| CN115147701A (en) | Infrared polarization image quality evaluation method and acquisition system | |
| Wu et al. | Image quantification of high-throughput tissue microarray | |
| Rodenacker et al. | Identification and quantification of phytoplankton by image analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190402 Termination date: 20190823 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |