determines the number of pancreas beta cells with ihc
TRANSCRIPT
i
DETERMINES THE NUMBER OF PANCREAS BETA CELLS WITH
IHC (IMMUNOHISTOCHEMISTRY) METHOD USING IMAGE
PROCESSING AND CALCULATING METHOD OF MATLAB
Oleh:
YOSPINA RERU
NIM : 642011801
TUGAS AKHIR
Diajukan kepada Program Studi Fisika, Fakultas Sains dan Matematika
guna memenuhi sebagian dari persyaratan untuk mencapai gelar Sarjana Sains
Program Studi Fisika
FAKULTAS SAINS DAN MATEMATIKA
UNIVERSITAS KRISTEN SATYA WACANA
SALATIGA
2016
Proceedings of International Symposium on BNCT
The Application of Nuclear Technology to Support National Sustainable Development:
Health, Agriculture, Energy, Industry and Environment October 26–28, 2015, Satya Wacana Christian University, Indonesia
Original paper available at http://... pp. …–…
Determines the number of pancreas beta cells with IHC (Immunohitochemistry) method using image processing and calculating
method of matlab
Yospina Reru, Jodelin Muninggar, Suryasatrya Trihandaru.
Departement of Physics, Satya Wacana Christian University, Salatiga 50711, Indonesia
Email : [email protected]
Abstract
Cell nucleus is located in the cytoplasm. The cell nucleus of pancreatic beta cells
located in the islets of Langerhans. The pancreas consists of exocrine and endocrine
glands. Endocrine function produces a rich part of the insulin, glucagon, and
pancreatic polypeptide. Insulin is a hormone that work to increase cellular energy
reserve which will affect the working system of the pancreas. Through the image of the
IHC method (Immunohystochemistry), can determine the amount of the cell nucleus in
the cytoplasm. The purpose of this project is to determine the number of pancreatic beta
cell nucleus contained in the cytoplasm using matlab. The calculation method of the cell
nucleus consists of 4 parts : determine color boxplot and operating boxplot RGB color
to black and white, imfill and noise reduction of image and the last step is labeling and
counting cell. Result of calculating the number of cell nucleus manually is 73, while the
calculation using matlab is 64 cell nucleus. It shows the results of less than 10%, which
means that this method could be used in determining the amount of the cell nucleus with
some improvement. However calculation using this method still has the blind spot that
each cell nucleus that is very close to the nucleus of cells that would otherwise be
counted as one nucleus, the cell nucleus which is close to the boundary would not be
counted, and this calculation is also affected by noise in the picture.
Keywords : Cell nucleus, Cytoplasm, Pancreas, IHC, Matlab.
Introduction
Cell nucleus is the main part of a cell which act to response to environmental
stimulation. Cell nucleus controls the entire cell activities and located in the cytoplasm,
in this case is the pancreatic beta cells which is located in islands of Langerhans.
Pancreas consists of exocrine and endocrine glands. Region of the endocrine has
function to produce insulin, glucagon, and pancreas polypeptide which are closely
related to the energy regulation in the body (Junqueira and Carneiro, 1992).
The pancreas’s endocrine gland composed of the Langerhans islands which is a
cluster that spread along the pancreas exocrine glands. Endocrine unit is referred to as
the island of Langerhans has 4 kind of cell i.e. alpha cells, beta cells, delta cells, and the
p. … Yospina Reru, Jodelin Muninggar, Suryasatrya Trihandaru.
cells on the pancreas polypeptide (Seungbum et al., 2007). The beta cells of the
pancreas on animals and humans are responsible for producing insulin. Insulin is
released into the blood by pancreatic beta cells vesicle in response to increase blood
sugar levels. Amount of beta cells can be an indicator for the levels of insulin, so that
when beta cells are detected more show insulin production will be more. If there is a
damage to the beta cells, it will effect to insulin levels. Damage of pancreatic beta cells
causes the body can not produce insulin, causing increase of blood glucose levels
(occurs state of hyperglycemic). The condition of hyperglycemic can result in the
formation of reactive oxygen species (ROS). Excessive ROS can cause oxidative stress
and may effect damage of pancreatic beta cells (Robertson et al., 2003).
Insulin has an important role in glucose metabolism. Insulin would help out
glucose entry into cells to do metabolism. Insulin is a hormone that function to increase
the cellular energy reserves that will affect the work system of pancreas (Erwin et al.,
2013). Beta cells damage occurs in diabetes mellitus. Thus, the amount of insulin will
be reduced. (Groop, 1999, 2001; De Fronzo, 1983; Masharani and Karam, 2001).
Through the image of the IHC (Imunohistochemistry) method, we can determine
the total of cells nucleus in the cytoplasm. However image from IHC visible irregular
cell shape and cell membranes that limit it not clear. It is difficult to determine the
boundary of the cell and calculate cell nucleus (Pezoa et al., 2015).
In determining clinical disease using the IHC method, required marker protein in
the cell nucleus, cytoplasm or membrane. The response expression of the cell determine
the result of this response will identify the presence of protein. Cell response against the
IHC method will be indicated by the expression of the color. Protein detection method
using marker (biomarker), also called IHC techniques. The underlying principle of this
technique is the staining of biopsy sample/tissue with specific antibodies for molecular
markers. IHC techniques is a method of staining on the tissue with the visualization of
antibody-antigen reaction using a secondary antibody conjugated to an enzyme, such as
peroxidase enzyme that catalyzes a reaction produces a brown color (Varghese et al.,
2014). Method of staining can show 4 types of beta cells on the island of Langerhans
cells i.e. alpha cells, beta cells, delta cells, and PP (pancreas polypeptide) (Ganong,
1995; Paulsen, 2000).
To help clarify the boundaries between cells tissues, image processing is needed.
Image processing allows to identify and obtain a clearer and accurate visualization of
Determines the number of pancreas beta cells with IHC (Immunohitochemistry) method
using image processing and calculating method of matlab. p. …
the tissue or cell (Ermatita et al. 2010). MATLAB was used in this study.
This project report to help determination the boundary between cells that makes
it easy to determine the amount of pancreatic beta cells nucleus on IHC slides using
MATLAB.
Materials and methods
Step 1. Color Boxplot Boxplot used to see the distribution of the data which can compare a lot of data
(more than 1 data). Usually to change the format of RGB to grayscale to see the
intensity difference, carried out the operation:
(1)
with:
(2)
(3)
Bloxplot is basis of the calculation with matrix and vector, if the data in the
matrix then there is one box per column and if the data in the vector, the data is only one
box. Bloxplot function to find a differentiator of color. In this section are differentiated
is gray, blue, and brown.
This section, format of two colors (black and white) can be made by selecting
the limit value of the intensity of use thresholding. For this case, it has always sought
pancreatic cells are in the middle of the image of the brown, so that the process of
conversion from RGB to grayscale is not done but required certain limits that should be
seen by a matrix R, G, and B separately. It is necessary for the calculation of the color
of each matrix. This can be done by using a boxplot. Boxplot will be obtained from the
limit value for conversion RGB into a two color.
Step 2. Imfill Color
This step including to the region where the filling is based on a number of
dilation, complementation and intersection. Toolbox provided is imfill. BW2 = imfill
(x) is an operation that is appropriate in this case because it displays the image on the
screen and lets us determine the area to fill with selecting points interactively using the
mouse. The area chosen is then replaced with a filling background black color so as to
distinguish between the nucleus and the cytoplasm more clearly.
Step 3. Noise Reduction This section to remove or reduce noise in the image of step before. Matlab used
black white area open (bwareaopen) which will remove all components connected to the
pixels that are less than the binary image, producing a binary image of the other.
Step 4. Labeling and Calculate cell nucleus
p. … Yospina Reru, Jodelin Muninggar, Suryasatrya Trihandaru.
Last step of this project is providing labeling and count the number in each cell
nucleus using regionprops in matlab. To determine the number of objects contained in
the image detected, utilizing regionprops function by looping the number of objects that
have been known.
Results and Discussion
IHC methods of insulin antibody used to detect the presence of insulin in the
pancreatic beta cell cytoplasm. Dark color (brown) in the cytoplasm appears on beta
cells showed the presence of insulin (Ridwan et al., 2012) (figure 1). The more a lot of
beta cell cytoplasm brown and calculated the cell nucleus, it can be estimated insulin
levels in these cells. Number of beta cell nucleus can be an indicator for the level of
insulin produced.
Figure 1. One of original images of the pancreatic beta IHC slides.
Process of digital image processing is obtained boxplot different RGB colors.
Significant differences were found in section B (figure 2). This means that the basic
color combination of digital image (RGB) can be operated IHC slides into black and
white, making it easier to classify between the nucleus and the cytoplasm. White color
outside the nucleus of the cell (cytoplasm) and background black color outside the
cytoplasm, are made equal to one color. Thus, what appears is the black color of the cell
nucleus in the cytoplasm (figure 3).
Determines the number of pancreas beta cells with IHC (Immunohitochemistry) method
using image processing and calculating method of matlab. p. …
Figure 2. RGB of gray, blue, and brown.
After that, the results of the black color made white through infill on black, so black
color (the cell nucleus) that are in the white area could be clearer and specific (figure
4).
Figure 3. Operation RGB color to black and white with
(a) black in white as in the cell nucleus, (b) beyond the
black and white as a background and (c) white as the
cytoplasm.
Figure 4. Output imfill black to white
Result of the calculation shows the cell nucleus results may varies. Number of cell
nucleus in the calculation using MATLAB is 64, can be seen in figure 5. While the
calculation using the manual method is 73 (figure 6).
p. … Yospina Reru, Jodelin Muninggar, Suryasatrya Trihandaru.
Figure 5. Output of using matlab calculation, 64 cell
nucleus.
Figure 6. Output of manual calculation, 73 cell nucleus.
Blind spot in this project is the nucleus of the cell that is located close to the
boundary would not be detected / counted and the cell nucleus that is very close to the
core that would otherwise be counted as one cell nucleus. In addition, the number of cell
nuclei are counted affected by noise settings, causing the calculation of the number of
cell nuclei also changed.
Conclusion and Remarks
In a related issue, calculation result is less than 10%, which means that the
method in this project could be used to determine the number of pancreatic beta cell
nucleus.
Acknowledgment
Author would like to thank lecturer of Faculty Science and Mathematics, Satya
Wacana Christian University, Mr. Suryasatriya Trihandaru and Mrs. Jodelin Muninggar.
Determines the number of pancreas beta cells with IHC (Immunohitochemistry) method
using image processing and calculating method of matlab. p. …
References
Ermatita, Hartati,S., Wardoyo, R., dan Harjoko, A. (2010). Medical imaging untuk analisis ekspresi gen.
Jurusan Ilmu Komputer dan Elektronika Fakultas MIPA Universitas Gadjah Mada. Indonesia.
Erwin, Etriwati, Muttaqien, Pangestiningsih, T.W., dan Widyarini,S. Ekspresi insulin pada pancreas
mencit (Mus musculus) yang diinduki dengan Streptozotocin berulang. (2013). Jurnal
Kedokteran Hewan Universitas Gadja Mada vol.7 No.2.;Indonesia.
Ganong, W.F. (1995). Fisiologi kedokteran. 14th
ed. Penerbit Buku Kedokteran EGC; 313-314. Indonesia.
Groop LC. (2001). Type 2 diabetes mellitus: Pathogenesis and Treatment. In Endocrinology and
Metabolism. : pp 607- 14, England: Mc Graw-Hill.
Junqueira, K.E. and Carneiro, J. (1992). Histologi dasar. Cetakan III. Penerbit Buku Kedokteran EGC;
430-435. Indonesia.
Masharani U, Karam JH. (2001). Pancreatic hormones & diabetes mellitus. In basic & clinical
endocrinology. 6th ed. Greenspan FS, Gardner DG (eds), : pp. 623-48. New York: Mc Graw
Hill.
Paulsen, D.F. (2000). Histologi and cell biology examination and board review. Cetakan I. McGaw-Hill
Publisher; 270-271, Singapore.
Pezoa,R., Rojas,R., Salinas,L., Pizarro,L., Reyes,J., and Gonzales,R. (2015). Segmentation of IHC-
stained breast tissues image using SVM. Center off Technological Innovation in HPC (CTI-
HPC) Universidad Tecnica Federico. Santa Maria.
Robertson, R.P., J. Harmon, P.O. Tran, Y. Tanaka and H. Takahashi. (2003). Glucose toxicity in beta-
cells: type 2 diabetes, good radicals gone bad, and the glutathione connection. Diabetes 52: 581-
587.
Seungbum, K., S. Jun-Seop, K. Hyung-Jung, K.C. Fisher, L. Mi-Ji and K. Chan-Wha. (2007).
Streptozotocin-induced diabetes can be reversed by hepatic oval cell activation through hepatic
transdifferentiation and pancreatic islet regeneration. Lab.Invertigation 87: 702-712.
Varghese,F., Bukhari,A.B., Abhijit De, and Malhotra,R. IHC profiler : an open source plugin for the
quantitative evaluation and automated scoring of immunohistochemistry image of human tissue
samples. (2014). Open Source Plugin for IHC image Scoring, India.
p. … Yospina Reru, Jodelin Muninggar, Suryasatrya Trihandaru.