Michael D. Torno DSc, Presenter: Nothing to Disclose

Nicholas P. Gruszauskas PhD, Abstract Co-Author: Nothing to Disclose

Roger Engelmann MS, Abstract Co-Author: License agreement, Hologic, Inc License agreement, General Electric Company License agreement, Toshiba Corporation License agreement, Deus Technologies, LLC License agreement, Riverain Technologies, LLC Research Grant, Riverain Technologies, LLC License agreement, MEDIAN Technologies License agreement, Mitsubishi Corporation License agreement, Quantitative Insights, Inc

Adam Starkey, Abstract Co-Author: Nothing to Disclose

Paul J. Chang MD, Abstract Co-Author: Co-founder, Stentor/Koninklijke Philips Electronics NV Technical Advisory Board, Amirsys, Inc Medical Advisory Board, Koninklijke Philips Electronics NV Medical Advisory Board, MModal Inc Medical Advisory Board, lifeIMAGE Medical Advisory Board, Merge Healthcare Incorporated

Samuel George Armato PhD, Abstract Co-Author: Nothing to Disclose

Software which de-identifies and indexes clinical data for a queryable research database was created. Users have the ability to save radiology reports and request de-identified medical images via the system’s web interface. Additionally, the opt-out paradigm provided a substantial number of consented patients and maximized the amount of data available to researchers. Autonomous operation of our dedicated research system resulted in minimal PACS performance degradation.

A system to mine, organize, anonymize, and request de-identified images from a radiological database was required to fulfill the needs of biomedical researchers. The system must function autonomously from a clinical PACS to minimize its impact on performance during clinical use. Independent operation allows keyword queries of anonymized radiology reports through a web interface; this interface also functions as a database creation and de-identified image request system. A custom database interface was designed to fill this need.

The software includes (1) Perl and VB apps to extract data from a clinical PACS and anonymize PHI in accordance with IRB and HIPAA standards; (2) an indexing search engine that allows keyword queries via a web browser; (3) PHP-based exporting of queried radiology reports with an option to request associated de-identified images through the Human Imaging Research Office at our institution. An “opt-out” IRB paradigm was created: outpatients in radiology reception areas are presented with an opt-out form to establish consent for use of clinical images and associated data for research.

The “opt-out” paradigm began in October 2008; to date 128,000 patients are enrolled and de-identified image data is available for query via our database interface. 1,324 patients have declined the study resulting in a 1% withdrawal rate. Previous “opt-in” paradigms resulted in an enrollment of less than 3,000 patients over a 5 year period. Over 1.2 million radiology reports encompassing over a decade of data were anonymized and indexed from our PACS and are available for use in medical research.

Torno, M, Gruszauskas, N, Engelmann, R, Starkey, A, Chang, P, Armato, S, An Anonymized Radiological Database with Open-source Search Engine and Image Request System for Biomedical Researchers.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13017417.html