Machine Learning Classification of Histologic Information of RNAscope Stained Breast Cancer Tissues Using Whole Slide Images
University of Canterbury
Time & Place
Mon, 30 Nov 2020 15:00:26 NZDT in E16 - Engineering Core
The genetic characteristics of breast cancers determine the mechanism that the cancer uses to grow, and therefore, the treatment options that will be effective. RNAscope in situ hybridization allows for the staining of particular RNA sequences in tissue samples, and therefore the quantification of gene expression. However, accurately quantifying the level of RNAscope staining in tumour and non-tumour tissue is extremely time-consuming when done manually. We aim to firstly evaluate existing solutions, and then to create a robust and fast method to segment important breast cancer tissue characteristics and quantify the level of RNAscope staining in relation to the different tissue types present. This will allow for easier and more objective appraisal of RNAscope stained breast cancer tissues.
Andrew Davidson is a PhD student in the Department of Computer Science and Software Engineering, supervised by Professor Ramakrishnan Mukundan. His work focuses on image processing and machine learning. His PhD started this year and is in the research proposal stage.