Student Name: Shrief Abdelazeez
Summary:
Shrief Abdelazeez is a Biomedical Engineer eager to apply his experience in this field to contribute to the community and industry. He earned his Bachelor's degree from the Faculty of Engineering at Cairo University, specializing in System and Biomedical Engineering, with an excellent cumulative grade and honors, achieving a GPA of 3.9. Subsequently, he completed his pre-master subjects with distinction, maintaining a GPA of 4.0. He then obtained his Master's degree in Biomedical Engineering from the Faculty of Engineering at Cairo University. The title of his master's thesis is "A Transfer Learning Based Framework for Multi-class Classification of Breast Cancer Using Whole Slide Images." His research interests encompass various areas, including Machine/Deep Learning Algorithms, Artificial Intelligence (AI) Algorithms, Statistical Analysis, Mathematical Modeling, Digital Signal Processing (DSP) Algorithms, Image Processing Algorithms, and High-Performance Computing (HPC). He is eager to utilize his expertise and skills to make meaningful contributions to the field of biomedical engineering.
Objectives:
The main objective of this project is to develop suitable automated methods for the recognition and quantification of tumourinfiltratin lymphocytes (TILs) in the tumour microenvironment, which has been recognised as a prognostic factor of relapse in BC, particular in triple-negative. Data for this investigation (images and annotation done by pathologists) coming from the other beneficent work– DC4-IISPV. At first, the studied methods should be able to imitate current TILs manual evaluation guidelines, based on the evaluation of the percentage of area or a number lymphocytes infiltrating stroma. Such evaluation is the only one possible manually, but when using digital slides, further quantitative variables could be identified that might provide additional information (e.g., density of infiltration). While the main target are hematoxylin-eosin slides (HE) (always available), lymphocyte-specific markers using immunohistochemistry will be also investigated to verify or extend observation in HE. In order to reach the main objective, a number of secondary aims should be investigated, as specifics tasks: 1) to develop colour standardization methods for the whole slides images (WSI) to homogenise the analysis of the images independently of the digital scanner where the image is obtained; 2) to develop a method for lymphocyte recognition using AI techniques, e.g., ML and DL algorithms; 3) using the lymphocyte map, development of a method for recognising infiltrated areas - TIL ; 4) selection of features with prognostic value (infiltrated area, lymphocyte density, node degree on the map, average arc length, and other parameters of TIL map measures).
Outcomes:
The main result will be one or more image-based markers that will provide prognostic information on BC relapse. The outcomes of this project are: 1) To obtain a colour standardization method for WSI with tissue stained with HE - in relation to task 1; 2) System for maps of TIL generation which based on AI techniques – in relation to tasks 2 and 3; 3) Automated assessment of TIL using HE or HE supported by immunohistochemistry using selected features – in relation to task 4; 4) Finding of correlates of TIL pattern in WSI with the BC relapse; and finally, all above, gives better understanding of the overall prognostic value and TILs-related prognostic factors.