ESTRO 2025 - Abstract Book

S2395

Interdisciplinary – Other

ESTRO 2025

Results A first AIDAVA prototype was developed to realize semi-automatic health data curation. Twelve tables with 184 data items were extracted from the EHR data sources. Data categories included demographic information, diagnosis, medical history, laboratory test results, medical imaging, surgical records, chemotherapy and radiotherapy. A personal health knowledge graph for breast cancer was constructed. Outputs were saved into RDF triples and stored in a graph database. Data query can be performed by using the SPARQL language to develop a local breast cancer registry. Conclusion The proposed workflow was successful in constructing knowledge graphs for patient-level data harmonization. Furthermore, it provides a potential solution to enhance medical data autonomy and facilitate clinical management in breast cancer.