Image Analysis for developing Quantitative Imaging Biomarkers
Wiro Niessen
Erasmus MC / Delft University of Technology

Abstract:

Biomarkers comprise anatomic, physiologic, metabolic, biochemical, biophysical, and molecular parameters which can be used to determine the presence and state of disease. Conventional biomarkers are commonly assessed by laboratory tests or physical examination. Owing to the advancement of imaging technology, and the development of contrast agents which are targeted to processes associated with specific diseases, the role of imaging biomarkers is rapidly increasing in medicine. Imaging biomarkers are expected to have a large impact in medicine, e.g. for better understanding the mechanisms of disease, for drug discovery and development, screening, improved diagnosis, and monitoring of treatment. In drug discovery, it has the potential to be used as a surrogate endpoint in clinical studies. In screening and diagnosis, it may lead to more sensitive and specific diagnosis. In treatment monitoring, it may be used to assess the response to treatment at an earlier stage, enabling individualized patient treatment. Whereas most conventional biomarkers (laboratory tests) are quantitative, a processing step is required to extract quantitative parameters from imaging data. Therefore, there is an important role for quantitative image analysis in the development of imaging biomarkers, and in their application in medical decision making and treatment monitoring. In this talk we will discuss advanced medical image analysis methods to develop, implement and validate quantitative imaging biomarkers. Examples of their potential will be shown in cardiovascular disease, neurodegenerative disease, and cancer.