Free Web-Based Application Designed to Streamline Medical Image Interpretation

Daniel Rubin, MD, an assistant professor of radiology, and his team initially designed ePAD in response to an unmet need in cancer imaging, but he says the tool can be used more generally quantitatively evaluate images and characterize disease. He said, “Currently, images are recorded in narrative text form. However, a narrative is a very muddy picture if a clinician and/or patient is trying to understand how the picture has changed over time and determine the response of a disease treatment. But if a radiologist is looking at images and all the data from earlier studies, such as dates and abnormalities, is contained in a table and a graph shows the changes in time, then it is easier for referring clinicians to understand and for computers to process.”

The other distinctive feature is that ePAD runs in a web browser. The advantage of doing this is the platform can be run anywhere, without needing to install software locally, or require a costly workstation, such as uses in radiology departments.

Dr. Rubin is currently in the process of establishing a pilot project of the system at the Stanford Cancer Institute. As part of the project, ePAD will be used to assess treatment success for patients who are matched to clinical trials using a smart database. Clinicians and researchers at other institutions have also begun using the tool, and Dr. Rubin has plans to increase its reach to create a huge, searchable medical image database. “We’re very excited about ePAD because we think it has far reaching implications,” he said.

Users, after downloading the platform, can utilize its graphic interface to review images, record semantic annotations, and make measurements. Information is stored in compliance with standards developed by the US National Cancer Institute’s Annotation and Image Markup (AIM), so that physicians can use the application to access other related imaging studies even if they were not interpreted using ePAD. Furthermore, annotations can be saved in a variety of formats so content can be simply and automatically searched across medical records systems, hospital image archives and the Semantic Web.

Related Links:
Stanford University