Level: Beginner
【Presenter】
Ryoji Chiba, Medical Director, Department of Pathology, Iwate Prefectural Central Hospital
[Co-presenter]
Chiba Ei
[Announcement summary]
Presentation title: As a spatial data and morphological analysis tool JMP -Upper respiratory tract carcinoma in situ and its precursor lesions using shortest distance cluster analysis Grading Attempt
Most malignant tumors of the upper respiratory tract mucosa such as the oral cavity, pharynx, and larynx are squamous cell carcinomas that occur in a layer of cells called stratified squamous epithelium that covers the surface of the mucosa. Furthermore, it is clinically known that this pre-stage or early stage lesion is found as a white to red phase of the mucosa, and epithelial dysplasia epithelial dysplasia, respectively, by observing microscopic specimens of tissues taken from patients. Carcinoma in situ is named. Furthermore, dysplasia is divided into three grades, from mild to mild mild, moderate moderate, and severe severe, depending on the abnormal cell morphology and the proportion of them in the epithelial layer. This discrimination is intuitively performed by a pathologist's inspection, and it is considered that it has some reproducibility, but there are not many objective studies. This time, we quantified the arrangement of cells (nuclei) in the epithelial layer and tried to investigate the difference between non-neoplastic (normal), dysplasia, and carcinoma in situ. The coordinates of the center of gravity of the cell nucleus are extracted on the micrograph by digital image analysis, and the minimum spanning trees (MST) connecting the centers of gravity are generated using the hierarchical cluster analysis of the shortest distance method of JMP ver.15. , When the histogram of the length of each branch was compared and examined, a difference was found between each group.
[Presenter Profile]
Ryoji Chiba
Former lecturer at Tohoku University (Aging Medical Research Institute, Pathological Organ Construction Research Field). medical doctor. Former overseas researcher (medicine), Ministry of Education (Aarhus University, Kingdom of Denmark).
In the field of human pathology, we have studied the pathological development process of diseases using quantitative morphology applying geometric probability theory and integral geometry, digital image analysis, and multivariate statistical analysis. There are research papers on early cancers occurring in liver cirrhosis, alveolar epithelium, pancreatic duct epithelium and endometrium, their precursor lesions, and liver metastasis of cancer.
( https://pubmed.ncbi.nlm.nih.gov/7804428/ , https://pubmed.ncbi.nlm.nih.gov/7804429/ , https://pubmed.ncbi.nlm.nih.gov/8402446/ , https://pubmed.ncbi.nlm.nih.gov/8135625/ , https://pubmed.ncbi.nlm.nih.gov/7840839/ , https://pubmed.ncbi.nlm.nih.gov/10560494/)
I am interested in mathematical methods that can be applied to their analysis from the viewpoint of cancer development process and histological diagnosis, and I am particularly interested in numerical classification methods such as cluster analysis and discriminant analysis. As a platform for statistical analysis, I am a JMP user from version 8 focusing on excellent data table functions and a flexible analysis environment via Fortran statistical subroutines on the mainframe, SPSS and SYSTAT on the PC.
Chiba Ei
Graduated from Future University-Hakodate, Faculty of Systems and Information Science, Department of Complex Systems Intelligence. While attending school, he was interested in complex system phenomena in physicochemical reactions, and conducted experiments and research on the mechanism of capsule formation in the alcoholic liquid phase of amino acid thermal polymers. Currently, I am also interested in digital image analysis, data science, and recognition and classification of morphology and images using neural networks.
