Feature Extraction and Evaluation of Colon Cancer Using PCA, LDA and Gene Expression

Abstract

At about 90 per cent of cases occurring in 50 people or older, colorectal cancer is the second most common type of cancer globally. Cancers may penetrate the deeper layers of colon and spread to other organs from the inner surface or from the muscular layer of the colon. The illness is lethal left untreated. Current screening of colon cancer is performed by colonoscopy. Visual examination of the colon and rectum mucosal lining with a photograph mounted on an endoscope is required. Testing will then biopsy suspicious areas. Images of colon (top photo) and normal colon tissue were found in the PR-OCT imagery. The green boxes display the probability values of the predicted "teeth" tissue patterns. Effective screening of colon cancer by versatile colonoscopy is carried out. The procedure involves visual inspection with a camera mounted on an endoscope of the mucus lining of the colon and rectum. Testing is then performed on suspicious regions. While this is the present standard of care, it has its drawbacks. This procedure relies first of all on visual identification, but the naked eye is difficult to detect minor lesions, and early malignancies also fail. Furthermore, only shifts in the bowels surface can be observed by visible endoscopy, not in its deeper layers.

Keywords: Colon Tissue, PCA, DNA, Colonoscopy, Machine Learning

Cite this Article: G. S. Pradeep Ghantasala, Bonam Tanuja, Golla Surya Teja, Anagurthi Sai Abhilash. Feature Extraction and Evaluation of Colon Cancer Using PCA, LDA and Gene Expression. International Journal of Software Computing and Testing. 2020; 6(1): 26–32p.