Semantic Search over DICOM Repositories

The integration of semantic representation in medical information systems has been seen as a key opportunity to achieve full interoperability between distinct technical solutions and to extract knowledge from existing repositories. In this paper, we propose an ontology-based medical imaging archive that provides a generic, dynamic and standard architecture to interrogate the repository. Moreover, the solution flexibility makes possible to improve the search results only by updating the ontologies used, without any changes in the software implementation. This archive is based on a triple store that follows the Rad Lex ontology. To validate the proposed system, the performance of storage and search queries were compared against the results obtained in a traditional PACS archive.