Widely spread breast cancer takes patients to an early grave. Early detection and the ability to predict the effectiveness of treatments are among the means used to fight various malignant cancers. Data federation from multiple data sources is needed for data analytics. Granularity, internality, structure and other characteristics often differ in federated data. We discuss alternative approaches to data federation and their theoretical basis, especially the stance regarding the ontology of data. We also develop an artifact to federate cancer data at a university hospital. Our approach and the artifact improved data interoperability, i.e., data federation. We suggest the same to be repeated in other data federation contexts.