Wake Forest University Imaging Informatics

About our group:

The Wake Forest University Imaging Informatics group consists of a number of faculty members and staff from Radiology, Biomedical Engineering, and Biomedical Informatics. We are involved in multiple large scale, multi-institutional cardiovascular and neurological imaging studies. One of our interests is to develop an integrated platform for effective and efficient management of the imaging and analytical results in these studies. We are very pleased that we are able to put together such a platform by integrating and extending existing open source technologies from various NIH-funded and privately formed projects. A core component of this platform is the XNAT system with some custom modules to handle analysis workflow and transfer of results from various cardiac imaging analysis systems. We are currently in the process of enhancing this platform to make it available for the research community via the Cardiovascular Research Grid (CVRG) project.

Our workflow:

The typical workflow for one of our projects is as follows:

Clinical scanning site performs imaging session and sends scan to Wake Forest University through our customized deployments of RSNA's MIRC or CTP.
As images are transferred through CTP, they are automatically deidentified (if needed), and tagged with the project and session name to help XNAT archive.
The images are routed to one of our DCM4CHEE PACS installations which will automatically forward to XNAT, a DICOM receiver that writes the images to a shared drive, and any DICOM enabled workstations or tools needed for that specific project.
XNAT auto-archives the scans in the corresponding project.
Protocol/scan check may be performed either automatically using a combination of rule engine and pipeline or by a data coordinator.
Analysts consult XNAT for scans that have not yet been analyzed and perform the analysis designated for each project. A typical project includes four or five types of analyses performed on different software platforms.
Once it has been performed, the analysis software writes a results file to a network folder accessible by XNAT.
The analyst selects a custom upload button in XNAT to trigger the pickup of the analysis data file. XNAT will find it in the network folder, parses and displays it on a form. The analyst verifies the new data and submits it to the XNAT database.
The analytical result data is now available for browsing and downloading through XNAT's interface.

About our XNAT project(s):

We use XNAT on a variety of cardiovascular as well as neurology imaging projects:

Diabetes Heart Study (DHS Mind, AADHS Mind)
Jackson Heart Study
CARDIA
MESA
Cardiovascular Research Grid
Predict
GENOA
Healthy-Normals
START
INFINITE
Football head impact study (HITS, KIDS, more info)
Plus several smaller studies including non-human primate imaging

We hold in XNAT over 9000 imaging sessions in 25 different projects. To store the results data from the different types of analysis we have added custom data types to XNAT plus we have developed mechanisms to upload and parse the results files.

Our goals for the XNAT Workshop:

We would like to understand new features and new directions of XNAT and discuss possible areas for collaborative development.