Arpad Bischof MD, Presenter: Employee, IMAGE Information Systems

Joerg Barkhausen MD, Abstract Co-Author: Nothing to Disclose

DICOM interoperability How to solve interoperability issues

The goal of this project was to normalize DICOM data to allow completely vendor independent hanging protocols. We collected 1.67 Million datasets from over 800 imaging centers in multiple countries. We analyzed the collected data and created a summary of the real world listing the variability of vendors and manufacturers in coding information in DICOM. We used artificial intelligence and peer review by radiologists and radiographers to interpret all data. We performed a comprehensive literature research about terminologies in medical imaging and developed a complex rule set to convert data into a vendor neutral archive format. We chose the newest publications of the DICOM committee or the IHE as the target formats. The system was able to convert 91.8% of data into a vendor neutral format. This approach will help manage DICOM data from multiple vendors and sources in a single network more efficiently, e.g. for global hanging protocol management. The system cannot create information which does not exist, e.g. completely missing information. System training for other geographical regions is pending. Further studies are required to assess the effects.

http://abstract.rsna.org/uploads/2014/14003889/14003889_qxaf.pdf

Bischof, A, Barkhausen, J, A New Approach for Semantic DICOM Interoperability.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14003889.html