Oralcancerdiagnostictechniques-1.docx

Evaluation of various Oral Cancer Diagnostic Technique Dr.G. Ulaganathan., M.D.S, Dr.K.Thanvir Mohamed Niazi M.D.S, Dr.Soundarya Srinivasan M.D.S, Dr.V.R.Balaji, M.D.S, Dr.D.Manikandan, M.D.S, K.A. Shahul Hameed ME, A. Banumathi.ME.Ph.D. Abstract Oral cancers are one of the most commonly occurring malignant tumors in the head and neck regions with high incident rate and mortality rate in the developed countries than in the developing countries. Generally the survival rate of cancer patients may increase when diagnosed at early stage, followed by prompt treatment and therapy. Recently, cancer diagnosis and therapy design for a specific cancer patient has been performed with the advanced computer aided techniques. The responses of the cancer therapy could be continuously monitored to ensure the effectiveness of the treatment process that hardly requires diagnostic result as quick as possible to improve the quality and patient care. This paper gives an overview of oral cancer occurrence, different types, and various diagnostic techniques. In addition, a brief introduction is given to various stages of immuno-analysis including tissue image preparation, whole slide imaging and microscopic image analysis. Keywords: Histopathology, Biomarker, Immunohistochemistry, Microscopic image analysis, Whole-slide imaging 1. Introduction Oral cancer is a frequently occurring and is the sixth most common cancer worldwide. Oral cancer is considered to be one of the most challenging malignancies of the head and neck regions that reportedly affect more than half million people every year worldwide [1] [2]. Especially, oral squamous cell carcinoma (OSCC) accounts for eighty-five percentages of all oral cancer which affects its site or origin and can spread to cervical node, lungs, liver and bones. During the year 2012, three million oral cancer patients (both sexes) have been identified out of which 145,000 cases were fatal. The incident rates of oral cancer are higher in the developed countries than in the developing countries. In India, the incident rate of oral cancer is 12.6 % per one million populations and remains the most common cancer in other Asian countries. In addition, the incident rate remains higher in some of the developed countries namely Australia, New Zealand, Germany, Poland, Denmark, Scotland and the USA. The five-year survival rate of early stage oral cancer patients is approximately 82%, and the advanced stage is about 20%. However, the five-year survival rate increased in developed countries such as the USA during the period from 1983 to 2006. At the same time, a negative trend is noticed in some countries like Brazil, Egypt, Japan, UK and Netherland where the number of causality has increased due to oral cancer [1].

Figure 1. Different types of potentially malignant oral lesions (a) carcinoma of tongue, (b) erythroplasia, (c) leukoplakia, (d) proliferative verrucous leukoplakia, (e) sublingual keratosis, and (f) candidal leukoplakia The main risk factors associated with the development of oral cancer and potentially malignant lesions (PML) are smoking, consuming alcohol and betel quid [3]. Early diagnosis of OSCC followed by prompt treatment can offer the best chance for cure and improve the survival rate of patients. Unlike others, the malignant portion of the oral cavity can be easily analyzed visually and/or different types of noninvasive method can also be used as a tool to in- crease the accuracy of physical examination. In spite of all, detection of OSCC is still an important and challenging problem for clinicians due to the fact that many lesions are asymptomatic at an early stage, and so regular screening of patients is mandatory. Generally clinically visible PMLs over oral cavity regions may become OSCC at later stage. Figure 1 shows few potentially malignant oral lesions namely, (a) carcinoma of tongue, (b) erythroplasia, (c) leukoplakia, (d) proliferative verrucous leukoplakia, (e) sublingual keratosis, and (f) candidal leukoplakia [4]. 





Squamous cell carcinoma of tongue (SCC) (Figure 1 (a)): This is more aggressive malignant tumor than conventional SCC appears in other regions of our body. The main cause for this tumor is unknown, however, the factors such as smoking, radiation, exposure to arsenic and coal tar may influence a lot for its development. Erythroplasia looks as bright red velvety plaques as displayed in Figure 1 (b), which is rarely occurred and clinically difficult to characterize. The main causes are presumed to be similar to the causes of SCC, and also mostly found in elderly people. Leukoplakia appears like thick, white patches on gum, inside of the cheeks, bottom of the mouth and sometimes on tongue also. These patches cannot be easily scraped off. Generally, most leukoplakia symptoms are benign, a few cases may become malignant, hence, it is not too dangerous, but sometimes may become so serious. Therefore, a patient with any one of these symptoms must

consult a dentist for further treatment. Figure 1 (c) to (f) show some types of leukoplakia. Early diagnosis of oral cancer and treatment aims at increasing the survival rate of a patient [5]. The standard method of revealing the oral lesion is benign or PML or OSCC by conventional oral examination, followed by histopathological analysis of biopsy taken from suspicious oral regions. In this process, pathologist use histopathological slides of biopsy samples taken from a patient, examine them under a microscope and make judgments based on clinico-pathological acumen [6]. The assessments are made only by visual examination of the deviation in the cell structure, shapes, tissue distribution, and grade of cancer. Hence, the results are purely qualitative and greatly suffered due to uncertain specificity and sensitivity [7][8]. Moreover, the awareness of endangerment of oral cancer among patients and also common people increases the demand for highly sophisticated and precise diagnostic techniques, which can detect the malignancy at an early stage. In order to meet these challenges, computer-aided diagnosis techniques has been introduced that utilizes the advanced image processing and machine learning techniques to produce quantitative diagnostic outcomes. 2. Different types of diagnostic techniques The requirements of oral cancer diagnostic techniques have been increased in the recent past in order to screen out malignant portion of oral lesion. Several diagnostic tools are currently available in the market. Among the most commonly used techniques are: histopathological study, vital staining, brush biopsy, and optical techniques. 

 



Histopathological examination is considered as golden standard test for cancer diagnosis. The effectiveness of the outcome depending on several factors including localization of suspected region and number of biopsy specimen taken from the malignant and normal tissue region [9]. Vital staining techniques are used to clinically highlight the malignant areas of oral lesion. It may help to acquire biopsy samples but unfortunately these techniques are not reliable. The brush biopsy techniques uses a tiny nylon brush to collect cytology samples which are scanned and analyzed by computer software in order to recognize whether individual cells are cancerous or not. Pathologists further examine the abnormal cells and positive results are attained only by a conventional incisional biopsy and histology examination. Incisional and excisional biopsies removing the part of the lesion with the inclusion of normal tissue is incisional biopsy. An attempt to clear the entire lesion is excisional biopsy. The possibility is that if the lesion is curable by local removal an excisional biopsy is recommended

The Biopsy was taken in the Department of OMFS, C.S.I College Of Dental Science and Research, Madurai

Generally, the abnormalities of cellular and molecular level in the tissues are not visible to human eye. It can only be observed by passing light through the tissues and detecting the changes in the optical spectrum. Optical spectroscopy techniques are commonly employed to locate malignant site of oral lesion prior to biopsy and histology investigation at early stages. 
The biomarker investigations became popular after the introduction of cellular and molecular studies about abnormalities in protein expression that have shown promising results in early diagnosis of oral cancer [10] [11]. The most popular and predictive biomarkers related to oral cancer development thus far reportedly available are TSG p53, TSG p16, Ki-67 antigen and DNA ploidy [4] [10]. For example p53 protein is regarded as an early event in carcinogenesis of OSCC. The p53 protein has been called as the guardian of genome, due to its ability in maintaining genomic stability, cell cycle control and apoptosis. In normal tissues, p53 protein has a short half-life and the amount of expression is very small which is normally undetectable. Conversely, the mutant p53 protein has prolonged half-life. It can remain in the tissues longer and can accumulate in cell nuclei to the levels that are detectable using specific staining process. Hence, p53 protein expression deserves particular attention in the case of early detection of oral cancer. Immunohistochemical (IHC) staining can detect the presence of the mutant p53 gene and it encodes protein with prolonged half-life, which induces overexpression of p53 protein in the nuclei. The p53 expression is considered to be positive, when the cell nucleus is detected with strong brown color which is shown in Figure 2. At present, these biomarker studies are adjuncts to help the routine histological examination, and their effective usages are still not commissioned due to the initial investment cost, lack of awareness, nature of complexity and limited number of studies reported up to date.

Figure 2: P53immunostained tissue image of OSCC. (a) 5x magnification; (b) Exported region(0.6mm); (c) Zoomed version of this slide under 40x in which positive (brown) and negative (blue) staining of p53 protein cell nuclei are marked by P & N respectively.

3. Microscopic image analysis Over the last decade, the nature of diagnostic methods and healthcare systems have been changing rapidly due to the vast availability of cancer patient data for diagnosis [12] [13]. Generally, classical diagnosis procedures are restricted to few diagnostic outcomes which mainly describe the stage/grade of the cancer or presence/absence of biomarkers such as HER2, p53, ki-67, estrogen receptor, and progesterone receptor [14] [15] [16]. The pathologists are well trained to extract the information from the tissue section that would help the oncologist to design the best therapy for a patient [8]. And also, these data could be used to understand the molecular characteristic of cancer with the view to improving therapy for a specific cancer patient.
Microscopic imaging technologies have been emerging in medical field to extract clinical and functional information from tissues and molecules [17]. Onco- pathologists are highly dependent on high-resolution microscopic images to assess the presence and the activity of target antigen in the tissues. In situ evaluation of the specific protein activation provides critical information about the status of tumor and it helps to design targeted therapy for a cancer patient. IHC is the most powerful tool in this field that uses marked antibodies to relate specific proteins in situ where the markers are located at sub-cellular regions namely nucleus, cytoplasm, and membrane. The evaluation of stains at these locations provides information about the nature and malignancy level of cancers. IHC techniques are introduced during early seventies for the assessment of tissue slides that have been given more importance in clinical research field after the introduction of high throughput imaging devices, and gained more attention as it has the capability to provide qualitative as well as quantitative diagnostic outcome [18]. The robustness of the outcomes is highly dependent on quantitative measurement rather than qualitative analysis of tissue slides, since manual outcomes are more subjective and suffers due to observers variability. The reliability of manual evaluation of tissue images is naturally subjective, heavily questioned and suffers due to inter/intra observer variability. In particular, modern pathology demands personalized medicine and therapy design for a

specific cancer patient that requires the extraction of information about protein activity at cellular level which cannot be easily obtained via visual examination [19]. Therefore, IHC analysis through computer-aided techniques is highly recommended to measure quantitative information from tissue under study. 4. Requirements of automatic techniques for IHC analysis Generally pathologists have been working strenuously on evaluating benign nature of tissue sections and their efforts can be used to analyze more complex or difficult suspicious tissue sections of tumors. For example, 20% of tissue sections are reportedly positive out of 1 million prostate biopsies performed every year in the USA, and the similar scenarios are found in other countries [20]. With the recent advancement in microscopic digital scanner, the whole tissue slides can be digitized and stored in memory devices. And the availability of image analysis algorithms with the large set of features, traditional way of diagnosis techniques could be replaced by digital pathologist [21]. Therefore, the requirements of computer-aided diagnosis (CAD) systems for medical image analysis have become increasing as it minimizes the workload on pathologist in terms of the time consumption and processing cost. For the past several decades, medical image analysis in cytology and radiology has been active research field that resulted in numerous products pertaining to a specific disease diagnosis [22] [23]. However, the imaging modalities, image features and automated procedures of these systems are notably different from histopathology analysis that cannot be directly applied to extract information from tissue section. Nevertheless, histopathology related problems attracts re- searchers working in the field of medicine, biochemistry and engineering due to tremendous development and production of microscopic scanners and aided software and their support in biomedical field [24]. In general histology images consist of more number of objects such as cell nuclei, stroma, lymphocytes, etc. which are widely distributed over the entire tissue sections and each needs to be accounted during diagnosis process.

Figure 3: Examples of different types of medical images (a) radiology, (b) cytology, & (c) histopathology. In the case of radiology, only a very few organs in the images are examined for

malignancy that usually located at a predictable place and mostly these images contain gray pixels. At the same time, cytology images look similar to histopathology images in terms of cell nuclei distribution; however, both images are taken at different magnitude levels for analysis, i.e., cytology images generally acquired at higher magnification level than histology images in order to capture in-depth information about internal cell nuclei structures. Figure 3 shows a small portion of (a) radiology, (b) cytology [24] and (c) histology images, in which histology images are more complex than other two due to the factors such as overlapping cell nuclei, blurred cell nuclei boundaries by noise and poor contrast. Nevertheless, these three automated systems have common sequence of steps such as segmentation, feature extraction, and classification. Although, there are several factors, which hinder the development of CAD systems for histopathology image analysis, among the most common problems include: i) complexity and diversity of tissue architectures are entirely different for each cancer tissue image, hence, it is very difficult to develop a universal CAD system for tissue analysis; ii) the algorithms developed for other medical image analysis cannot be easily adopted for histopathology; iii) no standard ground truth is available for the evaluation of CAD systems; hence, the performance outcomes are either fundamentally subjective or with minimal confidence testing. In spite of all these circumstances, the unworkable quantitative measurement on tissue images demands for the development of CAD systems and their important towards minimizing the workload of the experts. Generally, automated immunohistochemical (IHC) image analysis system consist of two major subsystems: 1) tissue section preparation and image acquisition; 2) automatic tissue image analysis, both involve a complex hardware and software elements [18]. Preparation of tissue section and image acquisition is a complex process that involves sequence of chemical and mechanical procedures with highly sophisticated hardware components. A brief introduction about various steps involved in immunostaining of oral cancer tissue images, tissue section preparation and image acquisition using whole slide imaging scanners are given in next sections.

Figure 4: IHC procedure from biopsy specimen to microscopic evaluation 5. Tissue section preparation and image acquisition IHC is a technique widely employed in histopathology to detect targeted protein expression in tissue sections by the use of labelled antibodies through antigen-antibody interaction. In particular, the detection is achieved by staining the tissues with the labelled antibodies, which have been selectively bound to the antigen under investigation. Professors A.H. Coons, H.J. Creech and R.N. Jones from department of Bacteriology and Immunology, Harvard Medical School and the Chemical Laboratory, Harvard University, introduced this technique during 1941 to detect bacteria and later it became one of the most powerful tools in diagnostic pathology [25] [26]. In digital pathology, IHC procedure is often used for localizing target antigen and also to identify abnormal cells such as those found in cancer tissues. Prior to microscopic evaluation of tissue images, IHC procedure contains two main steps namely, preparation of tissue sections and immuno-staining. These two steps are carried out by means of a complex chemical and mechanical procedure, which are having a great influence on the quality of the tissue images and also the repeatability of IHC analysis outcomes. In general, IHC workflow consists of the following steps namely; fixation, accessioning, grossing, tissue processing and embedding, sectioning, staining and screening interpretation and archive (Figure 4) [18]. Initially a portion of tissue is surgically removed from the suspected area of tumor location of cancer patient for biopsy verification and then the following procedures are carried out to visualize the antigenantibody interaction.

Step 1: Surgically removed biopsy-specimen is immersed in a fixative at the pathology laboratory in order to ensure the preservation of cell morphology and tissue architecture. Here depending on the imaging modalities and analysis method, different fixatives (crosslinking, precipitant etc) or methods (immersion, heat fixation, etc) have been used. Step 2: Once the tissue sample arrived, the details are entered into laboratory information systems (LIS) in the accessioning room. In order to trace the status of each tissue sample, a barcode label is given with each specimen. Step 3: After fixation, the tissue specimen is visually examined during grossing step for suspicious area which again requires microscopic investigation and then, these regions are excised as tissue blocks. Step 4: Tissue block is processed using formalin and is dehydrated for making ultrathin sections before embedding in paraffin or plastic material. Step 5: Embedded tissues are sectioned into thin slices and placed in a glass slides. Each slide is given a unique barcode to retrieve the information about staining protocol for the particular section. Step 6: The main part of tissue preparation is staining procedure, which has been performed in direct or indirect way. Direct method is processed with only one labelled antibody, directly reacting with the antigen in the tissue sections. This method is rarely used due to poor signal amplification. Indirect methods are commonly employed in IHC staining; here, primary antibody reacting tissue section and a secondary antibody counterstained with primary antibody in order to visualize the distribution of nuclei architecture, resulting in aa higher sensitivity by greater signal amplification. Later, tissue sections are treated by dehydration, clearing, mounted and cover slipped. Step 7: The pathologist examines immuno-stained tissue sections under micro- scope for screening and interpretation.

Step 8: The outcomes are updated to the oncologist for designing an appropriate therapy for cancer patient.

Generally biomarkers i.e., proteins or target antigens are colored by differ- ent staining technique to visualize/highlight the positive/negative cell nuclei or region of interest from the background. In general, p53 positive cell nuclei are colored using Diaminobenzidine (DAB) staining which is one of the most commonly used stain and it gives dark brown color to biomarkers. And the hematoxyline (H) is used as a counterstain that gives blue color to the nega- tive cells and others. Therefore tissue sections of oral cancer are stained with H-DAB staining protocol to analyze p53 protein content for diagnosis [3]. 6. Whole slide imaging techniques:

Digital pathology is the field by which histopathology tissue slides are scanned to produce very high-resolution digital images using whole-slide imaging (WSI) devices [27] [19] [28]. These digitized images are given to automated image analysis tool for further processes such as segmentation, nuclei detection, feature extraction, feature selection, classification and scoring. In the past, most of the pathology labs have shown keen interest on digital processing and started their process using high resolution imaging techniques. This could possibly achieved only through the advancement in microscopic techniques as well as related hardware and software components. Lots of researchers are constantly working towards the development of advanced digital technologies in histopathology and relevant field that has made the possibility of widespread accessibility, long time and easy preservation of tissue slides with considerable amount of price. The growth in storage capacity, display technologies, processor speed, bandwidth expansion and networking allows WSI more effective and practically more feasible for extracting quantitative features from the tissue images which helps the pathologist to perform their work. However, the transformation of diagnostic pathology from traditional light microscopy to WSI technology has been slower as compared to radiology related technology due to various reasons including investment cost, memory capacity, storage of both glass slide as well as its high-resolution image. Wetzel and Gilbertson have developed the first CAD system based on high- resolution WSI imaging devices during 1999 (described in [29]). Since, the requirements for WSI in the field of diagnostic pathology have steadily increased to improve the reliability and reproducibility of diagnostic outcome [27] [30]. In General, WSI scanners consists of the following components: illumination systems, optical microscopic devices, a precise focusing equipment that places tissue slide image on a digital scanner, high-resolution display, storage unit and software and hardware components, which handle the sequence of operations performed during scanning process. The final output of WSI is a complete digital virtual slide which can be freely viewed at any standard range of magnification and also carry out the necessary operations by the experts. WSI technology provides promising results compared to traditional light microscope by avoiding several limiting factors such as poor image quality, low-resolution overview of tissue section, control of slide navigation, and the excessive amount of time requires to review a slide [31]. In addition, WSI systems can be easily accessed remotely by the pathologist at place on the globe to give their opinion about the tissue under study and also helpful for the people who are working in the area of both clinical as well as biomedical engineering to carry out teaching and research activities. 7. Conclusion Over the past few decades, an automatic histopathological image analysis has shown tremendous growth in terms of both hardware and software with the help of advanced computing techniques. Many researchers have been working in this field aiming to develop efficient as well as reliable automatic techniques for diagnosing cancer that could help the oncologists in designing therapy for a specific patient. The survival rate of cancer patients could be increased only when the disease is diagnosed at early stage. However, cellular and molecular studies about biomarkers protein in tissue sections have shown promising results in diagnosis of cancer at early stage. In particular, biomarker

proteins such as p53, ki-67, p15 etc. are reportedly considered as the most popular predictive markers for early stage of oral cancer development. IHC techniques are generally employed to visualize the activation of these biomarkers by staining the tissue samples which are taken from diseased patients. The marked anti- bodies are located at sub-cellular locations, such as cell nuclei, membrane and cytoplasm, which needs to be analyzed carefully for extracting meaningful quantitative information. However, the presence and the status of the disease could be studied only through a high-resolution microscope. After the introduction of high-resolution WSI technologies, histology image analysis became popular and most of the clinical onco-pathological labs are showing keen interest on digital pathology related products. In this paper, we have given a brief introduction to oral cancer and different types of diagnostic techniques which include man- ual procedures currently employed and the requirements of automatic image analysis methods

References 1. G. Del Corso, A. Villa, A. Tarsitano, A. Gohel, Current trends in oral cancer: A systematic review, Cancer Cell & Microenvironment 3 (3). 2. S. Balakrishnan, S. Manoharan, L. M. Alias, M. R. Nirmal, Effect of cur- cumin and ferulic acid on modulation of expression pattern of p53 and bcl-2 proteins in 7, 12-dimethylbenz [] anthracene-induced hamster buccal pouch carcinogenesis. 3. S.-W. Chang, S. Abdul-Kareem, A. F. Merican, R. B. Zain, Oral cancer prognosis based on clinicopathologic and genomic markers using a hybrid of feature selection and machine learning methods, BMC bioinformatics 14 (1) (2013) 1. 4. C. Scully, J. V. Bagan, C. Hopper, J. B. Epstein, et al., Oral cancer: current and future diagnostic techniques, Am J Dent 21 (4) (2008) 199–209. 5. R. Mehrotra, S. Yadav, et al., Oral squamous cell carcinoma: etiology, pathogenesis and prognostic value of genomic alterations, Indian journal of cancer 43 (2) (2006) 60. 6. C. D. Fletcher, Diagnostic Histopathology of Tumors: 2-Volume Set with CDROMs, Elsevier Health Sciences, 2007. 7. M. A. Gavrielides, B. D. Gallas, P. Lenz, A. Badano, S. M. Hewitt, Ob- server variability in the interpretation of her2/neu immunohistochemical expression with unaided and computer-aided digital microscopy, Archives of pathology & laboratory medicine 135 (2) (2011) 233–242. 8. S. Di Cataldo, E. Ficarra, E. Macii, Computer-aided techniques for chro- mogenic immunohistochemistry: status and directions, Computers in biol- ogy and medicine 42 (10) (2012) 1012–1025. 9. J. A. Woolgar, Histopathological prognosticators in oral and oropharyngeal squamous cell carcinoma, Oral oncology 42 (3) (2006) 229–239. 10. H. Polanska, M. Raudenska, J. Gumulec, M. Sztalmachova, V. Adam, R. Kizek, M. Masarik, Clinical significance of head and neck squamous cell cancer biomarkers, Oral oncology 50 (3) (2014) 168–177.

11. U. Swaminathan, E. Joshua, U. K. Rao, K. Ranganathan, Expression of p53 and cyclin d1 in oral squamous cell carcinoma and normal mucosa: An immunohistochemical study, Journal of oral and maxillofacial pathology: JOMFP 16 (2) (2012) 172–177. 12. J. Rittscher, R. Machiraju, S. T. Wong, Microscopic image analysis for life science applications, Artech House, 2008. 13. M. N. Gurcan, L. E. Boucheron, A. Can, A. Madabhushi, N. M. Rajpoot, B. Yener, Histopathological image analysis: a review, Biomedical Engineer- ing, IEEE Reviews in 2 (2009) 147–171. 14. H. Masmoudi, S. M. Hewitt, N. Petrick, K. J. Myers, M. A. Gavrielides, Automated quantitative assessment of her-2/neu immunohistochemical expression in breast cancer, IEEE transactions on medical imaging 28 (6) (2009) 916–925. 15. C. Primkhajeepong, P. Phukpattaranont, S. Limsiroratana, P. Boonyaphiphat, K. Kayasut3, Evaluation of color based breast cancer cell images analysis, Sonklanakarin Journal of Science and Technology 32 (3) (2010) 231. 16. A. Law, K. Lam, F. K. Lam, T. K. Wong, J. L. Poon, F. H. Chan, Image analysis system for assessment of immunohistochemically stained prolifer- ative marker (mib-1) in oesophageal squamous cell carcinoma, Computer methods and programs in biomedicine 70 (1) (2003) 37–45. 17. F. Ghaznavi, A. Evans, A. Madabhushi, M. Feldman, Digital imaging in pathology: whole-slide imaging and beyond, Annual Review of Pathology: Mechanisms of Disease 8 (2013) 331–359. 18. R. L. Taylor CR, Education guide: immunohistochemical staining methods, Sixth edition, Dako North America, 2013. 19. A. Madabhushi, Digital pathology image analysis: opportunities and chal- lenges, Imaging in Medicine 1 (1) (2009) 7–10. 20. J.-W. Song, J.-H. Lee, J.-H. Choi, S.-J. Chun, Automatic differential diag- nosis of pancreatic serous and mucinous cystadenomas based on morpho- logical features, Computers in biology and medicine 43 (1) (2013) 1–15. 21. A. Madabhushi, S. Viswanath, G. Lee, P. Tiwari, Medical image informatics for personalized medicine, Critical Values 6 (3) (2013) 30–33. 22. W. Schroeder, L. Ng, J. Cates, et al., The itk software guide, The Insight Consortium. 23. M. R. Lamprecht, D. M. Sabatini, A. E. Carpenter, et al., Cellprofiler free, versatile software for automated biological image analysis, Biotechniques 42 (1) (2007) 71. 24. L. He, L. R. Long, S. Antani, G. R. Thoma, Histology image analysis for carcinoma detection and grading, Computer methods and programs in biomedicine 107 (3) (2012) 538–556. 25. A. H. Coons, H. J. Creech, R. N. Jones, Immunological properties of an antibody containing a fluorescent group., Experimental Biology and Medicine 47 (2) (1941) 200–202. 26. A. H. Coons, M. H. Kaplan, Localization of antigen in tissue cells ii. improvements in a method for the detection of antigen by means of fluorescent antibody, The Journal of experimental medicine 91 (1) (1950) 1–13.

27. G. Begelman, M. Pechuk, E. Rivlin, M. Sabo, et al., A microscopic telepathology system for multiresolution computer-aided diagnostics, Jour- nal of Multimedia 1 (7) (2006) 40–48. 28. G. Bueno, R. Gonzalez, O. D ́eniz, M. Garc ́ıa-Rojo, J. Gonzalez-Garcia, M. Fern ́andez-Carrobles, N. V ́allez, J. Salido, A parallel solution for high resolution histological image analysis, Computer methods and programs in biomedicine 108 (1) (2012) 388–401. 29. J. Ho, A. V. Parwani, D. M. Jukic, Y. Yagi, L. Anthony, J. R. Gilbertson, Use of whole slide imaging in surgical pathology quality assurance: design and pilot validation studies, Human pathology 37 (3) (2006) 322–331. 30. E. A. Krupinski, A. R. Graham, R. S. Weinstein, Characterizing the development of visual search expertise in pathology residents viewing whole slide images, Human pathology 44 (3) (2013) 357–364. 31. S. Kothari, J. H. Phan, T. H. Stokes, M. D. Wang, Pathology imaging informatics for quantitative analysis of whole-slide images, Journal of the American Medical Informatics Association 20 (6) (2013) 1099–1108. 32. Robert.E.Marx, DDS ., Diane Sterne, DDS.,MS., Oral and Maxillofacial pathology : A Rationale for Diagnosis and Treatment volume 1-second edition Chapter 1 , Page 1