Martin Huber PhD, Presenter: Employee, Siemens AG

Juergen Fritsch PhD, Abstract Co-Author: Employee, Multimodal Technologies, Inc

Gerhard Kohl PhD, Abstract Co-Author: Employee, Siemens AG

Daniel L. Rubin MD, Abstract Co-Author: Grant, General Electric Company

We developed a method to link image annotation and measurement with automated reporting of quantitative imaging results. This approach may improve report quality and workflow in reporting quantitative image findings.

As radiology becomes quantitative, there is a need to convey increasing amounts of this quantitative information in reports. It can be time consuming and error-prone to accurately convey all measurements, locations, and other findings, and a report may inadvertently omit information. Our goal was to develop an approach to automate the reporting of quantitative imaging results to improve reporting accuracy, to streamline report generation, and to check report completeness.

We developed a reporting solution that pre-populates a structured radiology report with quantitative measurements made in a commercial imaging workstation. The radiologist completes the report using conventional voice-driven dictation, adding findings and impressions using macros or narrative dictation. All information is filled into a structured reporting template using the HL7 Clinical Document Architecture (CDA) making use of RadLex and the Annotation and Image Markup (AIM) standard of caBIG to semantically annotate image findings like lesions (lesion name, anatomic location, and measurement). The template also summarizes the measurements on the lesions reported in the prior examination, providing feedback to the radiologist useful to avoid inadvertent omissions in reporting lesion measurements. Further, the system displays warnings based on severity distinguishing e.g. mandatory and recommended content.

Semantic annotation of reporting content enables automatically pre-populating a structured radiology report in CDA format with quantitative imaging results, which may improve report quality and the reporting workflow. Generating the report as the user annotates images provides feedback wherein the user can quickly see the lesions measured previously and be prompted as to which to measure on the current exam. We will expand on this functionality to examine the structured and codified report content to determine if other quality criteria are met such as consistency with prior reports.

Huber, M, Fritsch, J, Kohl, G, Rubin, D, Automated Structured Reporting of Quantitative Imaging Results.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034222.html