Research project

Synopsis of research project and overall aims

My research project concern the advance image analysis of immunohistochemical images for biomarker validation and diagnosis in cancer research.
Nowadays, it is possible using high technology scanning hardware generate high resolution images of a tissue section or a whole tissue micro array. However, the size of images and the number of images scanned have also increased so developing solutions for dealing with digital slides handling and analysis is of great interest from commercial and academic point of view.

The general aim of my study is first to provide a platform of software utilities and algorithms for automated quantitative image analysis. Second, in long-term, I’m as well interested in developing classification models incorporating image data and clinical information with view towards separation of disease types and patient nature.

Objectives of research project

A recently purchased TMA scanner will be used. A first library of virtual TMA slides is already ready to be analysed in melanoma for a quantitative annotation and another one for biomarker quantitative expression analysis will soon be available.

The first TMA library is manually annotated by a pathologist and will be used for testing the performances of the automated analysis algorithms.

 The platform of software utilities and algorithms will be mainly focussed in the following key issues of TMA data analysis:

• TMA image de-arraying. This step includes first identification of each spot in the TMA image, as well all the missing spots and second quantifying the amount of tissue in each spot. 
• Integration of software tools for core and TMA data base management.
• Core annotations following staining levels
• Cytoplasm
• Membrane
• Nuclear
•  Identify malignant tissue from stromal tissue and non malignant tissue.