An eScience Program for the Advancement of Care in Oncology - Imaging as a Biomarker and Big Data Analysis 

The OncoSpace Project (JHU - Rad Onc)

The explosive increase in data volume associated with technological advances in detection and computation poses major challenges to the continuing advancement of health care.  A foremost hurdle lies in the utilization of multi-faceted data where the sheer volume impedes decision making and collaborative research. Equally disconcerting is the lack of an effective and efficient forum to promote learning from one another’s experience. In response, we propose to develop a fundamentally new healthcare informatics infra-structure and research paradigm where data are not sent away; instead they are maintained locally to support approved queries and analyses that are sent in by internal and external healthcare providers, researchers, trainees and patients. The inherent data sharing model enhances the efficiency of clinical research, fosters data reuse and increases the breadth of clinical knowledge available to the community. 

Incorporating Imaging Data into OncoSpace

Our long term goal is to maximize the availability and utilization of clinical data to improve healthcare delivery and reduce its cost.  Our technology is OncoSpace, a database for structured information modeled after the successful SkyServer program developed by the Department of Physics and Astronomy at Johns Hopkins University.  OncoSpace is intended for federation at different oncology centers to support data sharing.  We choose radiation oncology as a meaningful use-case model where multi-modality imaging and multi-faceted non-imaging data are employed in the management of a patient’s treatment. We have demonstrated the feasibility of web-based remote query of dosimetry and toxicity data in an OncoSpace installed at a collaborator site. While there are non-technical issues such as privacy protection that must be addressed, our proposed effort to the Toshiba’s initiative is to expand the capabilities of OncoSpace to incorporate multi-modality imaging for decision support. In particular, we will focus on the management of patients with head and neck (H&N) cancers where treatment and intervention strategies are significantly challenged with the increasing volume of repeat multi-modality (CT, PET, MRI and Ultrasound) imaging and where traditional data management has not been conducive to data-reuse and decision support.  Radiation treatment of H&N cancer is highly suitable to explore imaging as bio-markers.