Chapter Eighteen - Biomedical image characterization and radiogenomics

Abstract

Much of the discussion of personalized medicine has been focused on molecular characterization of tissue samples using genomic and proteomic technologies. However, as lesions are spatially and temporally heterogeneous, these techniques are limited. They require biopsies or invasive surgeries to extract generally small portions of tumor tissue that do not allow for complete characterization of the lesion. On the contrary, imaging is noninvasive and has already been often repeated in routine treatment practice for clinical use. Imaging has great potential to guide therapy by providing a more comprehensive view of the entire lesion and can be used on an ongoing basis to monitor the development and progression of disease or its response to therapy. The purposes of radiogenomics research are usually threefold, each targeting different aspects of disease characterization. Here we consider cancer as an example. First, radiogenomics aims to correlate imaging characteristics (i.e., the imaging phenotype) with gene expression patterns, gene mutations, and other genome-related characteristics and is designed to facilitate a deeper understanding of tumor biology and capture intrinsic tumor heterogeneity. Second, radiomic features have been investigated for their power to indicate the genomic status of tumors, which exploits imaging techniques as potential noninvasive tools for probing cancer molecular biomarkers relieving patients from repetitive biopsies. Third, radiogenomics establishes the association between tumor phenotype and personalized medicine by integration of tumor radiomic and genomic features for disease diagnosis, prognosis, and treatment design/management. Due to the noninvasive nature of medical imaging and its ubiquitous use in clinical practice, the field of radiogenomics, despite its newest family of “omic,” is rapidly evolving, and the initial results have been encouraging in the past few years.