Towards Interoperable ECGs: Converting Proprietary XML to DICOM

Electrocardiograms (ECGs) are of central importance for the detection of various cardiovascular diseases. However, the storage and exchange of ECG data within and across healthcare institutions is hindered by limited interoperability. The "Advancing Cardiovascular Risk Identification with Structured Clinical Documentation and Biosignal Derived Phenotypes Synthesis" (ACRIBiS) project aims to provide a solution by developing a FAIR (Findable, Accessible, Interoperable, and Reusable) ECG data infrastructure across multiple German hospitals. This paper presents an open-source, general-purpose converter for transforming proprietary ECG XML files into the Digital Imaging and Communications in Medicine (DICOM) format, thereby ensuring a standardised data representation. In addition to ECG raw data, the converter processes metadata and waveform annotations using value codes from established medical terminologies (e.g. SCP-ECG). In this work, we evaluate the performance of the converter on 36,933 clinical ECGs stored within GE Healthcare Muse®. Results demonstrate high performance and compliance with DICOM standards. Future work will extend compatibility to other vendors' formats and integrate the tool into workflows at various ACRIBiS partners.