Past research projects

  • Context-sensitive Teleconsultation Infrastructure

    Context-sensitive Teleconsultation Infrastructure

    This research investigated "context-sensitive telemedicine." Methods were developed to automatically identify the most relevant image slices containing anatomical structures of interest, achieved by combining a corpus-based anatomy knowledge base with structure delineation through image registration, deformation, and atlas mapping. Through these methods, a "clinical compression" of large imaging studies can be achieved, with only key image slices selected and transmitte... More

  • Data Structuring & Visualization System for Neuro-oncology

    Data Structuring & Visualization System for Neuro-oncology

    An appropriate review of a patient's medical record often requires that a physician review multiple clinical documents while mentally noting issues related to the case. Physicians must review what the findings were, the chronology of events, spatial/temporal patterns of disease progression, and the effects of interventions. Among all this data, the physician must filter out information irrelevant to the current clinical context of care. The growing complexity of the medical record mak... More

  • DataServer: An XML-based Data Gateway for Clinical Information

    DataServer: An XML-based Data Gateway for Clinical Information

    This research detailed development of an electronic infrastructure that established a single access point for patient information using XML as the basis for representing queries and results to and from clinical data sources. The variety of current medical standards has resulted in a highly heterogeneous, distributed data environment, making it difficult to aggregate patient information into one view. The proposed system works on two levels: 1) intra-site, mapping data sources in a single i... More

  • An Engineering-based Approach to Individually Tailored Medicine

    An Engineering-based Approach to Individually Tailored Medicine

    This grant examined and developed a health plan process engineered according to an individual patient's needs - that is, individually tailored medicine. An agent-based electronic information infrastructure permits fast, reliable, intelligent access to distributed patient records. Natural language processing is employed to extract information from documents, permitting the automatic instantiation of a new type of "phenomenon-centric" data model in support of investigative exp... More

  • An Imaging-based Disease Model for Understanding Bone Health

    An Imaging-based Disease Model for Understanding Bone Health

    This project addressed the development of a disease model for bone development to better comprehend the genesis of osteoporosis, and specifically looks at modular methods for integrated visualization of clinical data. The model will be built by using quantitative magnetic resonance (qMR) to examine bone growth in young adults and assess trabecular bone quality. A Bayesian belief network (will be constructed using these qMR values with other imaging and observational data. Two patient popul... More

  • PACS for Evidence-based Medical Practice, Research & Education

    PACS for Evidence-based Medical Practice, Research & Education

    The broad, long-term objective of this Program Project Grant was to develop an effective imaging-based information and healthcare delivery system to support clinical practice, research, and education. The aims of the grant were to: 1) evolve PACS into an effective infrastructure that promotes the objectification of subjective patient clinical symptoms; 2) develop methods for improving the characterization of medical data through structured data collection, natural language processing of me... More

  • Towards ITM: Probabilistic Models of Cerebral Aneurysms

    Towards ITM: Probabilistic Models of Cerebral Aneurysms

    The objective of this proposal, which builds on previous MII research, is to create an informatics infrastructure to explain the genesis, progression, and treatment of intracranial aneurysms (ICAs). A set of processing pipelines will extract free-text and imaging clinical data for development of a phenomenon-centric data model (PCDM) and Bayesian belief network (BBN) to model ICAs. Together, the PCDM and BBN allow for progressive refinement of the ICA model as clinicians examine new cases.... More