FP27 Clinical and research data integration using a common coordinate framework for the gut

Introduction Common Coordinate Frameworks (CCF) have been developed to provide standardised reference maps for human anatomy. These maps enable data-location from the macroscopic organ level through to individual cell-types. CCFs are typically organ-specific and allow the recording and comparison of anatomy location-specific data across multiple individuals in spite of the natural variation of organ lengths and shapes. These frameworks have recently been developed and extended to support the Human Cell Atlas initiative but are also relevant for spatially indexing clinical and pathological data more generally. We have developed a CCF for the human gut including 1D, 2D and 3D models of the small and large intestines and are proposing the computational concept of Location Cards to facilitate an open software architecture for sharing of gut-related data sets across multiple sources.

Methods Abstracting the tube-like structure of the gut, we developed a linear (1D) model consisting of gut landmarks, such as ileo-caecal valve and hepatic or splenic flexures, and regions such as ileum and transverse colon. Regions of Interest are marked by start and end points along the linear model, where distances are captured either in absolute terms, e.g. 10mm distal from the splenic flexure, or in relative terms, e.g. one third along the ascending colon. Locations identified in 2D or 3D, such as from a patient’s CT scan, are mapped onto the linear model using computational algorithms. We have introduced the concept of Location Cards (LC) as a standardised representation of multi-modal anatomical, radiological and surgical location data. Existing and new biomedical research and clinical data can be annotated using LCs, and an open LC-based software architecture is proposed that can be deployed alongside existing systems.

Results Our 1D, 2D and 3D gut models are publicly accessible in multiple formats. A demonstrator application has been developed to show the utility and spatial cross-mapping of our gut models and the concept of Location Cards in support of linking gut data repositories. Specifically, we show how clinical data from our own project can be linked to the gut data section of the HuBMAP reference atlas as well as others.

Conclusion Common Coordinate Frameworks in biomedical research have become an important computational tool. A CCF for the human gut is now also available and holds significant potential for the integration of research and clinical data related to the gut. The next step is to share the CCF widely, including the models as well as the software architectural description for Location Cards, and encourage its extended use to maximise the integration potential.