gms | German Medical Science

67. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS), 13. Jahreskongress der Technologie- und Methodenplattform für die vernetzte medizinische Forschung e. V. (TMF)

21.08. - 25.08.2022, online

Incremental extraction of DICOM header to a FHIR Database for Data Integration Centers of the Medical Infomatics Initative

Meeting Abstract

  • Jan Maluche - Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany
  • Ralf Lützkendorf - Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany
  • Johannes Bernarding - Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany
  • Christian Bruns - Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. 67. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS), 13. Jahreskongress der Technologie- und Methodenplattform für die vernetzte medizinische Forschung e.V. (TMF). sine loco [digital], 21.-25.08.2022. Düsseldorf: German Medical Science GMS Publishing House; 2022. DocAbstr. 195

doi: 10.3205/22gmds122, urn:nbn:de:0183-22gmds1224

Published: August 19, 2022

© 2022 Maluche et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at



Introduction: The MII core dataset defines basic and extension modules on how data are stored and made available [1]. The basic modules include information about patients, encounter, diagnoses, procedures and laboratory values. We enrich this dataset with additional information based on dicom tags, which are retrieved from a locally deployed XNAT instance [2]. This extended dataset can be used for better study recruitment and the building of cohorts. We introduce an incremental approach which extracts data from the XNAT api, transforms them to a FHIR resource and stores them in a FHIR based datawarehouse [3].

Methods: We developed a spring batch-based approach, defining readers, processors and writers for the data [4]. Calls to a RESTful interface are made from the reader to assemble information about the tree-like xnat internal structure (subjects, projects, experiments, scans). Based on XNAT internal modification timestamps, requests to the target FHIR server are made to determine if resources needs to be created, updated or are already up-to-date. The first two cases require more information about every image of the imaging sessions assigned to this experiment and subject. The dicomheader are retrieved and forwarded to the internal processor. The processor applies a modular mapping process which maps from dicom-tags to FHIR imagingstudy resource. This resource is forwarded to the writer, which stores the resource in a FHIR server.

Within the resources, an individual metadata extension is created. This FHIR extension stores a timestamp, according to internal XNAT data. There are two possible values provided by XNAT. At first, the timestamp when an experiment was added. At second, an internal last-modified timestamp is created, when any user edits are recognized. The latter is only defined, if the dataset was modified by the user. Therefore, it is important to interpret both possibilities and store the newest date within the extension. This metadata extension is interpreted in the reader process.

The hole approach is packed in a one-shot docker container, which is highly customizable through configuration files. It is planned to evaluate the whole process within the MII-project MIRACUM where results are expected soon.

Results: Just now, evaluation is planned and prepared. An internal dataset of 2 TB MRI-scans from 3 different scanners is used. For quality evaluation, a subset of this dataset is taken and processed. These results are compared to the results of linux4health/dicom-fhir-converter [5], which maps locally stored DICOM images to FHIR resources.

For speed evaluation, the whole dataset is processed and timings are taken.

Discussion: By now, only basic information about the presence of images are processed. By researcher's demand, the amount of information can be easily extended through the modular implementation. The deployed process can be easily included into an existing and site individual architecture.

Conclusion: In this manuscript about a basic development of a FHIR generator for a scientific DICOM archiving system (in this case XNAT) we showed a proof of principle to enrich the information in the FHIR based central datawarehouses of the data integration centers in the university hospitals by the headers of picture data.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.


Semler SC, Wissing F, Heyder R. German Medical Informatics Initiative. Methods Inf Med. 2018 May;57(S 1):e50.
Herrick R, Horton W, Olsen T, McKay M, Archie KA, Marcus DS. XNAT Central: Open sourcing imaging research data. NeuroImage. 2016 Jan 1;124, Part B:1093–6.
HL7 FHIR v4.0.1 [Internet]. [cited 2021 Sep 7]. Available from: External link
Spring Batch: Overview [Internet]. [cited 2021 Sep 7]. Available from: External link
dicom-fhir-converter [Internet]. Linux For Health; 2021 [cited 2021 Sep 7]. Available from: External link