The HiGHmed Medizininformatik‑Consortium project, funded by the German Cancer Research Center (DKFZ) under grant 01ZZ1802D, ran from 1 January 2018 to 30 April 2023. Its core objective was to create a cross‑institutional infrastructure for the exchange of therapeutic options for patients with hepato‑pancreato‑biliary (HPB) cancers, using a shared molecular tumor board and a data mart that would capture clinical experience and integrate evidence from the literature. The initiative was carried out in close collaboration with the university hospitals in Heidelberg, Göttingen, Hannover and, from 2019, Cologne, all members of the HiGHmed network. The consortium also involved the Molit GmbH for software selection and the German Cancer Research Center’s own data integration teams.
Technically, the project focused on comprehensive data modelling and integration. Patient data were modelled to include molecular genetic information from next‑generation sequencing, radiological and radiomics data, and clinical records of diagnosis and disease course. The resulting data archetypes were implemented in openEHR and made publicly available through the HiGHmed Clinical Knowledge Manager at https://ckm.highmed.org. These archetypes provide a foundation for future data extraction and exchange interfaces. In parallel, the consortium established a data mart that systematically stores care data for HPB patients, enabling systematic querying and the incorporation of “world knowledge” from published studies and databases.
A key deliverable was the establishment of a joint molecular tumor board that meets monthly across the four sites. Since its inception in 2020, the board has reviewed patients with atypical characteristics, using a standardized template for presenting findings. The board’s operation relies on a broad consent framework agreed upon by the Medizininformatik‑Initiative and the German Data Protection Conference, and on ethics approvals that cover HPB‑specific aspects. The legal and organisational groundwork allowed the exchange of patient data between the sites and the submission of research queries to the use and access committees at each university hospital.
The project also introduced the SnaP tool, a search engine for identifying similar patients within the MeDIC repositories. During the extended funding phase (1 January 2023–30 April 2023), SnaP was implemented and demonstrated using test data from the Hannover and Cologne sites. Although the tool’s performance metrics were not formally reported, its deployment marked a functional step toward real‑time similarity searches in clinical oncology.
Several planned concepts were revised during the project. An initial plan to create a dedicated omics data integration center (omicsDIC) at the DKFZ was abandoned due to personnel and organisational changes. Instead, a generic omicsDIC concept was developed, allowing each MeDIC to implement the necessary modules. For HPB patients, whole‑genome or exome sequencing was deemed impractical; therefore, panel sequencing was adopted, and its results were stored as molecular pathology findings in each site’s MeDIC using the shared data models. ETL pipelines were established to transfer data from source systems into the MeDICs.
The software required for managing the tumor board was evaluated, and VITU from Molit GmbH was selected. However, technical interface challenges and limited availability of developers meant that VITU was ultimately installed only at the University Hospital Cologne for internal use. This limitation highlighted the need for more robust, interoperable solutions in future iterations.
Overall, the HiGHmed project successfully produced a set of openEHR archetypes, established a functioning cross‑site molecular tumor board, and created a data mart that captures HPB patient care data. These achievements lay the groundwork for more advanced analytics, evidence‑based treatment recommendations, and broader collaboration across German university hospitals in precision oncology.