Gpf4Med: A large-scale graph processing system applied to the study of breast cancer

Today, there is much knowledge that is not exploited from the clinical records from thousands of patients treated at different centres. In part, this is because traditional databases fail from revealing undiscovered correlations that can contribute to improve clinical outcomes and to reduce the costs of patient care. This paper presents a new graph processing framework for clinical data, which can leverage from cloud computing to address large-scale studies. Also, a case study of breast cancer with relevance for the clinical practice is presented. This case is successfully addressed using a dataset consisting of 15,000 reports from 1,000 anonymised patients, demonstrating the capability of the framework for indexing and searching large, heterogeneous datasets.