CORBEL Report on final assessment and generalization of integration platform

Task 3.4 in CORBEL (“Data integration and management services for image-driven and genomicsdriven biomarker studies”) was designed to develop a common IT framework to support data handling and analysis for both clinical and preclinical biomarker research, across disease areas. The effectiveness of the selected approaches has been demonstrated using national and international research projects in this domain. It was investigated whether the data integration and imaging solutions implemented for the selected use cases are not only operational and used, but can also serve as best practices within the framework of shared services between the (ESFRI) Biological and Medical Research Infrastructures in the medical domain. This has been demonstrated with four independent use cases across the spectrum from preclinical to clinical biomarker research:

1. In the preclinical research use case tools were developed for easily extracting, importing, archiving preclinical studies based on magnetic resonance imaging (MRI) from several vendors, including tools for automated image processing. The herein developed workflow will be made available to the wider preclinical research community, hence allowing a simplified exchange and (re)use of image datasets among preclinical imaging centers.
2. In the PROOF [1] use case (osteoarthritis), a generic infrastructure was developed for integrative analysis of medical imaging, genetics and clinical data, supporting imaginggenetics association studies and advanced radiomics analyses using multivariate machinelearning to develop novel predictive markers. Best practices are given.
3. In the IMI APPROACH [2] use case (osteoarthritis), a generic infrastructure was implemented largely shared with the other two clinical use cases. This integration solution is operational and successfully used in the daily project workflow. The solution regarding the data processing and loading is set up to be generic and the final approach taken can be used as a best practice for future projects, which is further detailed in the recommendations towards the end of the present document.
4. For Cancer Core Europe [3], a collaboration project managed between a number of leading European cancer centres, a data sharing taskforce has been set up to implement the data integration solutions. Within this project a use case revolving around patients with nonbreast related tumours in conjunction with a BRCA1 or BRCA2 mutation, was selected to execute with support of CORBEL. The implementation of the integration solutions are discussed and conclusions and recommendations shared.

The core of the CORBEL common IT framework for genomics and imaging studies in biomarker research consists of the tranSMART (data integration and browsing platform), cBioportal (patientlevel genomics platform) and XNAT (image archive platform) solutions. The present document describes how this infrastructure facilitates multi-modal biomarker research. The four use cases highlight what is needed for multimodal biomarker research, such as data capture, data curation pipelines, standardization, transformation, integration, and analysis. We conclude with recommendations regarding practical use of the common IT framework for future projects.