Imaging Biomarker Ontology (IBO): A Biomedical Ontology to Annotate and Share Imaging Biomarker Data

Abstract

Imaging biomarkers refer to radiological measurements that characterize biological processes of imaged subjects and help clinicians particularly in the assessment of therapeutic responses and the early prediction of pathologies. Several imaging features (size of a lesion, volume of a tumor, blood perfusion in a specific anatomical region, anisotropic water diffusion in a particular tissue region, etc.) are quantified and reported in the clinical practice. The growth of the number of research studies addressing imaging biomarkers and the increasing use of these measurements in the radiological routine necessitates the use of semantic research tools. The use of semantic technologies will enable to efficiently retrieve imaging-related data and to enhance the interoperability in the biomedical field. While many efforts have been conducted regarding the definition of a standardized vocabulary to support the sharing of the imaging biomarker knowledge, the definition of the term “imaging biomarker” stills inconsistent. In this paper, we introduce our motivation for semantically describing this concept and we outline shortcomings of the state-of-the-art methods. Here, we propose a semantic representation of the imaging biomarker concept that is based on the articulation of its three main semantic axes, namely the measured quality, the measurement tool and the decision tool. The developed ontology is called the Imaging Biomarker Ontology (IBO) and uses existing biomedical ontologies. A preliminary use case is studied to illustrate the utility of IBO in annotating quantitative and qualitative imaging data from the TCGA (The Cancer Genome Atlas) collection.