Biomedical imaging and Information Technology

Finbarr O’Sullivan is the Principal Investigator of a number of research projects in Biomedical Imaging supported by Science Foundation Ireland (SFI 11/PI/1027), National Institute of Health and the National Cancer Institutes (NIH/NCI CA 65537 and NIH/NCI P01 CA-042045-23). The Biomedical Imaging team comprises of Jian Huang, Janet O’Sullivan, Eric Wolsztynski, Tian Mou, Fransisco Hernandez-Fernandez, Zhaoyan Xiu and Sarah Murphy.

The past decade has seen major advances in the technologies for quantitative image data capture in biology and medicine. Increasingly, researchers need to make use of (sequences of) volumetric image data in their work. However, local biologic variability of the image target as well as fundamental restrictions in resolution characteristics (limited by dose constraints) produce data that have significant error. As with a number of fields, advances in instrumentation and data generation capabilities have far out-paced the development of analysis tools that are capable of extracting useful information. This is especially the case with technologies used for in-vivo imaging studies.

Our work focuses on the need for effective and reliable techniques capable of making quantitative use of the complex information that in-vivo imaging offers. Our research concentrates on the development of statistical methodologies for application to such data and using the results for enhanced prognostic assessments. A range of analytic tools are useful in this work, including approximation, asymptotics analysis, function and functional inference, inverse theory, numerical optimization and stochastic modeling.

We identify two main areas of application of our research:

* PET Research in Cancer (funded by SFI, NIH, and NCI)
http://euclid.ucc.ie/pages/staff/osullivan/res-cancer.htm

This work relates to imaging of cancer and its response to treatment by surgery, chemotherapy and radiation. Data from on-going clinical studies at the Cork University Hospital as well as at a number of collaborating US universities are used for development and validation of techniques. We develop and implement statistical tools for the analysis of PET, PET/MR and PET/CT data mainly.

* Stroke/Alzheimers research (funded by SFI and NIH)
http://euclid.ucc.ie/pages/staff/osullivan/res-stroke.htm
 
The goal is to develop and test the clinical viability of cerebral perfusion neuro-imaging with MR,CT and PET for ischemic stroke patients and geriatric patients at risk for Alzheimer’s disease (AD). Data from a number of clinical centers are used, including from an Irish hospital setting.