Data and model aggregation for radiomics applications: Emerging trend and open challenges

Abstract

Radiomics is a quantitative approach to analyzing medical multi-layered images in combination with molecular, genetic and clinical information, which has evidenced very promising results especially in the field of oncology. Radiomics applications, however, pose several challenges from the computational viewpoint, and their effective deployment in real-world scenarios require to carefully put in place a number of sophisticated information fusion algorithms and approaches. The paper overviews some relevant works in this area, by depicting a clear picture of the current trends in information fusion methods defined at the level of the data sources as well as at the level of the models learned for diagnostic purposes. While the potential of radiomics to enhance established diagnostic, prognostic and even therapeutic approaches in different diseases merged rather clearly, the analysis evidences several issues that require to attract further attention in the research community.