Data sharing and data-driven processes are gaining significant traction in the healthcare research and industry. Medical data, such as magnetic resonance (MR) images from previous patients, are increasingly utilized to automate diagnostic procedures for future patients. Although there is substantial research on the use of MR images in data-driven processes, there are notable gaps in studies addressing how processes acquire fit-for-purpose data, particularly with respect to ensuring data anonymization, standardization, and fulfilment. This paper introduces an extended Capability Description Language (CDL+) to bridge the gaps in the integration of data and medical data-driven process.