Daniel L. Rubin MD, Presenter: Research grant, General Electric Company

Daniel Korenblum MS, BA, Abstract Co-Author: Nothing to Disclose

Vijaykalyan Yeluri MS, Abstract Co-Author: Employee, General Electric Company

Perry Frederick PhD, Abstract Co-Author: Employee, General Electric Company

Robert J. Herfkens MD, Abstract Co-Author: Nothing to Disclose

We have implemented the AIM image metadata standard in a commercial PACS workstation, a step toward translating this important technology into routine clinical and research practice. The workstation functionality and workflow are not significantly altered, and new applications based on AIM annotations can be deployed. We expect similar implementations in other workstations in the future.

The NCI caBIG project recently developed standards in image annotation and markup (AIM) for capturing and sharing quantitative and qualitative information in images. To date AIM is not implemented in the clinical PACS. Our goal was to implement AIM in a commercial PACS to demonstrate the feasibility and utility of broad adoption of this important standard in clinical imaging workstations.

We implemented the AIM image metadata standard in the GE RA1000 PACS workstation (GE Medical Systems, Milwaukee, WI). We extended the pre-existing annotation tools to save the annotation information in the AIM format. Text annotations created after drawing an ROI provide the label for the ROI. AIM objects are stored in a database separate from the images to facilitate image query and to reduce any performance impact on the operational PACS database. We evaluated our implementation by annotating images containing cancer lesions in serial CT studies and by validating the resulting AIM. We also created a software module to analyze AIM annotations to produce automated quantitative summaries of lesion measurements for tracking tumor burden.

Our implementation is transparent to the user; while annotating images in the workstation, an AIM document is produced simultaneously, without interfering with the speed of image annotation or workflow. AIM objects contain all pertinent image metadata, and they are interoperable with those obtained from other AIM-compliant systems. Our tumor burden tracking application consumes image annotations to summarize the lesion status in cancer patients. To our knowledge, ours is the first implementation of AIM in a commercial PACS workstation demonstrating collection of image annotations and their implementation in a clinically-useful application.

Rubin, D, Korenblum, D, Yeluri, V, Frederick, P, Herfkens, R, Semantic Annotation and Image Markup in a Commercial PACS Workstation.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9003853.html