An Annotation Approach for Radiology Reports Linking Clinical Text and Medical Images with Instructional Purposes

Abstract

Clinical annotation is a critical issue in multiple medical disciplines, and particularly within the radiology domain. The current research deals with a broad concept of annotation that can be used when working with the clinical textual descriptions from radiology reports as well as with the digital images that are attached to this kind of reports. In this context, the main goal of this paper consists of providing physicians and radiology staff with a technical structure to fully integrate the annotation processes both for clinical texts and their associated medical images. Unlike other annotation approaches, the present proposal primarily has an instructional purpose that allows practitioners to experiment with annotation mechanisms from an academic perspective. The approach proposed is complemented by the Clavy system to retrieve clinical corpora coming from multiple medical repositories and convert them into reconfigurable digital collections. Its application, together with the development of the CARS annotation plug-in, has enabled the instructional deployment of multimedia-based radiology collections, thus making them available as educational resources. A case study has been used to show the potential of this approach and its possibilities to annotate and analyze different examples of radiology reports.