Towards improved cancer diagnosis and prognosis using analysis of gene expression data and computer aided imaging.

With the increasing cost effectiveness of whole slide digital scanners, gene expression microarray and SNP technologies, tissue specimens can now be analyzed using sophisticated computer aided image and data analysis techniques for accurate diagnoses and identification of prognostic markers and potential targets for therapeutic intervention. Microarray analysis is routinely able to identify biomarkers correlated with survival and reveal pathways underlying pathogenesis and invasion. In this paper we describe how microarray profiling of tumor samples combined with simple but powerful methods of analysis can identify biologically distinct disease subclasses of breast cancer with distinct molecular signatures, differential recurrence rates and potentially, very different response to therapy. Image analysis methods are also rapidly finding application in the clinic, complementing the pathologist in quantitative, reproducible, detection, staging, and grading of disease. We will describe novel computerized image analysis techniques and machine learning tools for automated cancer detection from digitized histopathology and how they can be employed for disease diagnosis and prognosis for prostate and breast cancer.