Facilitating multi-disciplinary, knowledge-based support for breast cancer screening

Abstract

In order to increase the accuracy of breast cancer screening, the diagnostics of imaging modalities (X-ray, ultrasound or magnetic resonance (MR)) is assessed alongside results of histopathological or cytopathological studies as well as clinical information about a patient. This procedure is called 'triple assessment' in the UK. We have undertaken a project to develop a multiple ontology-based medical image-annotation and reasoning system to support this procedure. Our system integrates image-annotation tools for drawing, image analysis and feature extraction, with ontologies compliant with the proposed language standard, namely OWL. This integration is carried out by the mapping between images and instances in the ontologies. In this paper, we outline the ambit of the project in further detail and discuss some of the architecture and design issues that we have resolved in order to make our system modular and semantic web-enabled. The resulting knowledge base should not only support medical practitioners in the triple assessment process but also provide a resource from which personalised patient information could be delivered via the web direct to patient or care provider. Finally, we describe the progress that we have made thus far on the implementation of this system.