Jafi Alyssa Lipson MD, Presenter: Nothing to Disclose

Dawn Charisse Nwamuo MD, Abstract Co-Author: Nothing to Disclose

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

Accurate interpretation of breast imaging studies depends on adequate preparation, which in turn depends on the radiologist’s ability to integrate a myriad of data: clinical history, pathology, and results from prior imaging. This complex task is time-consuming and error-prone. We developed and evaluated an informatics tool to provide a visual summary of pertinent clinical information and tested the hypothesis that our tool will improve efficiency and completeness of preparation for breast imaging interpretation.

We developed an Annotated Breast Map (ABM), a tool that provides a visual display of data pertinent to comprehensive assessment of breast imaging studies. A patient-centric temporal data model populated by extracting data from electronic medical record and radiology information systems, the tool integrates demographics, breast cancer risk factors, and exam indication, and provides annotated images of both breasts, summarizing breast findings graphically, including location of prior breast procedures, pathologic diagnoses at each location, and correlates on other imaging modalities. Images are annotated using the caBIG Annotation and Image Markup (AIM) standard, enabling automated summarization of lesion identities and locations. We evaluated the ABM and tested our hypothesis by comparing preparation time and completeness of a breast imaging specialist who evaluated 10 breast imaging studies.

Preparation for interpretation required synthesizing 1) exam indication, 2) pathologic diagnoses, 3) prior exam findings, 4) clinical context, and 5) history provided by the patient.  The baseline preparation time of the radiologist without using ABM was 25 minutes/case on average (stdev 5 minutes; range 15-35 minutes). The baseline preparation completeness was 100% (5/5 pieces of pertinent data reported). Performance significantly improved by using the ABM, with preparation time decreasing to an average of 15 minutes/case (stdev 5 minutes; range 10-20 minutes; p < 0.05; CI 5.30 – 14.70) and completeness in preparation maintained at 100% (5/5 pieces of pertinent data reported; p < 0.05).

Our visual approach to information integration improved efficiency and maintained the accuracy of preparation for breast imaging interpretation.

Annotated Breast Maps that summarize pertinent clinical information could streamline and improve quality of breast imaging interpretation.

Lipson, J, Nwamuo, D, Rubin, D, The Annotated Breast Map: A Novel Paradigm to Improve Interpretation in Breast Imaging.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9013607.html