Decomposed storage model suitable for DICOM images

A DICOM decomposed storage model (DCMDSM) is suitable as a storage layer for projects where DICOM images are stored once and retrieved whenever necessary, according to a study published online Feb. 3 by the Journal of the American Medical Informatics Association.

Lead author Alexandre Savaris, of the Computer Science Department at the University of Kaiserslautern in Germany, and colleagues strived to reduce the time spent for query and retrieval workloads by evaluating a data model created to provide full-content storage for DICOM images and full-metadata indexing.

“Due to its simplicity, the DSM storage model is suitable for extensions and customizations…In this work, the original DSM architecture is adapted to incorporate characteristics found in DICOM image files, aiming to provide a full-content storage model with performance gains for query/retrieval operations,” wrote the authors.

The DCMDSM model proposed by Savaris et al differs from previous projects in that all standard and proprietary tags extracted from DICOM image files are stored and indexed to provide full flexibility on query construction and execution; metadata access through the DICOM standard can be performed with predicates and any unique, required, or optional search key; and content retrieval is possible at pixel data or full-content level.

To evaluate the storage model, the researchers performed experiments to observe its behavior and determine its suitability for storing heterogeneous DICOM images without schema modifications, querying metadata using single and multi criteria predicates, and retrieving full-content images. The same experiments were performed on a well-established DICOM archive and image manager for comparative analysis.

Though the proposed model was found to be 0.6 to 7.2 times slower in storing content in the established DICOM archive, it was 48 percent faster in querying individual tags. The DCMDSM was outperformed in querying groups with large numbers of tags and had low selectivity. On the other hand, performance gains reach up to 79.1 percent when the number of tags is balanced with better selectivity predicates. The proposed model was also 48.3 percent faster than the well-established model in executing full-content retrieval.

“As future work, it is proposed to evaluate the same model with large datasets (not necessarily heterogeneous), comparing results from the current single-node implementation with results from a version built over a distributed relational database management systems (RDBMS)/column-oriented database management systems (DBMS),” wrote the study’s authors.