Data miner retrieves cases for clinical comparison

April 30, 2006

A prototype system created by Canadian researchers uses data mining algorithms to automatically search, identify, and retrieve clinically relevant cases when studies are opened in a PACS.

A prototype system created by Canadian researchers uses data mining algorithms to automatically search, identify, and retrieve clinically relevant cases when studies are opened in a PACS.

Details of the system were presented Friday at the Society for Imaging Informatics in Medicine meeting. Designed by researchers at the St. Boniface General Hospital Research Center, in conjunction with the University of Manitoba, the system creates a classification data repository. Cases are correlated in the repository and are linked to existing reading workstations and RIS/PACS systems, said presenter Sergio Camorlinga, Ph.D.

The data mining tool is activated when a patient exam is selected and opened on a PACS work list. It searches the database and presents a list of correlated cases for review and comparison.

The prototype system is currently being used with mammography exams; additional applications are planned.