Innovating Medical Image Analysis via Spatial Logics

Current computer-assisted medical imaging for the planning of radiotherapy requires high-level mathematical and computational skills. These are often paired with the case-by-case integration of highly specialised technologies. The lack of modularity at the right level of abstraction in this field hinders research, collaboration and transfer of expertise among medical physicists, engineers and technicians. The longer term aim of the introduction of spatial logics and spatial model checking in medical imaging is to provide an open platform introducing declarative medical image analysis. This will provide domain experts with a convenient and very concise way to specify contouring and segmentation operations, grounded on the solid mathematical foundations of Topological Spatial Logics. We show preliminary results, obtained using the spatial model checker VoxLogicA, for the automatic identification of specific brain tissues in a healthy brain and we discuss a selection of challenges for spatial model checking for medical imaging.