DicomShield: A Pseudonymization Proxy for the Secondary Use of Imaging Data in the Research Context

Image data collected during clinical workflows is valuable for secondary use in research. Usually, these images are stored with identifying information about the patient in Picture Archiving and Communication Systems (PACS). Here, they are findable and retrievable using the Digital Imaging and Communications in Medicine (DICOM) standard by providing a patient's name or ID. However, research use is subject to strict data protection requirements. Identifying information such as names must be removed from the image metadata and IDs must be replaced with pseudonyms. In this work, we present DicomShield, an open-source proxy that can be placed in front of a DICOM-based PACS and can interact with external pseudonymization services to ensure consistency of pseudonyms across different source systems. DicomShield uses pynetdicom and FHIR pseudonymization interfaces. Experiments showed a 160 ms slowdown per image retrieval due to the "on-the-fly" pseudonymization, which is acceptable for most research purposes. It is available as open-source on GitHub https://github.com/imi-ms/DicomShield.