Text

Introduction: Understanding disease development increasingly relies on integrating diverse data sources, such as care records, cohort data, and genetic information. This integration reveals complex links between disease onset and progression, making the digital representation of multimodal data essential.

Genomic data plays a pivotal role in precision medicine [1], [2], especially for complex conditions like epilepsy [3]. Advances in gene discovery have significantly improved rare disease diagnostics. However, the growing number of known disease-associated genes also complicates interpretation.

Standardized data models and collaboration are key to effectively using genomic information in research and care. In oncology, molecular tumor boards [4] help tailor treatments through interdisciplinary collaboration. A similar approach is needed in neurology: molecular epilepsy boards can guide precision therapies based on the latest genetic insights.

As part of the BMBF-funded TreatION project, we developed a demonstrator for such a board.

Methods: Our approach begins with a requirements analysis to guide the integration of diverse, health-related data for precision medicine. This includes patient data, diagnostic measurements (e.g., MRI, EEG), and external sources. The goal is to design a flexible, scalable data model that allows for interchangeable outputs and avoids tool-specific dependencies.

Centralizing this multimodal data on secure hospital servers is key. Genetic variant information from research literature must also be stored for use in pathogenicity scoring. This structured integration supports advanced analytics, clinical decision-making, and research.

Relational databases, though scalable, are limited by rigid schemas. Alternatives like NoSQL or RDF offer flexibility but pose interoperability and usability challenges. A hybrid approach using relational databases in an Entity-Attribute-Value (EAV) [5] format provides the needed flexibility, but it must be supported by robust metadata to ensure data usability.

TreatION hospitals currently collect patient data using REDCap forms, which vary across sites. A key task is transforming these into a unified, adaptable format that supports future expansion.

Results: To enable rapid access to existing and future data, we implemented a hybrid EAV database model, avoiding table changes and allowing flexible schema extension. The database includes three main tables:

• Patient: Stores immutable demographic data (e.g., birthdate, sex, ethnicity), with data origin tracking.
• Clinical Findings: Uses an EAV format to record variable data (e.g., diagnoses, MRI, EEG, and genetic results). Each entry includes a parameter, value, coding system, and metadata (e.g., source, contributor). REDCap forms are mapped into this structure, with grouped fields linked by a cluster ID.
• Literature: Compiles curated knowledge from OMIM, ClinVar, and gnomAD in EAV format for reference during variant interpretation.

For visualization and exploration, we integrated the data model with cBioPortal, enabling dashboard-driven analysis of clinical, mutation, and literature data relevant to epilepsy.????

Conclusion: We developed a hybrid EAV-based data model for integrating heterogeneous clinical and genomic data in precision epilepsy medicine, demonstrated in the TreatION project. The model supports flexibility to incorporate new datasets, thereby seeding the effective use of the data, standardized mapping, and scalability, with future compatibility toward frameworks like OMOP CDM. Integration with cBioPortal enables effective visualization and analysis, laying the groundwork for interoperable and collaborative precision medicine tools.

The authors declare that they have no competing interests.

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

Literatur

[1] Tesi B, Boileau C, Boycott KM, Canaud G, Caulfield M, Choukair D, et al. Precision medicine in rare diseases: What is next? J Intern Med. 2023;294:397–412. DOI: 10.1111/joim.13655
[2] Knowles JK, Helbig I, Metcalf CS, Lubbers LS, Isom LL, Demarest S, et al. Precision medicine for genetic epilepsy on the horizon: Recent advances, present challenges, and suggestions for continued progress. Epilepsia. 2022;63:2461–75. DOI: 10.1111/epi.17332
[3] Symonds JD, Zuberi SM, Johnson MR. Advances in epilepsy gene discovery and implications for epilepsy diagnosis and treatment. Current Opinion in Neurology. 2017;30:193. DOI: 10.1097/WCO.0000000000000433
[4] Schwaederle M, Parker BA, Schwab RB, Fanta PT, Boles SG, Daniels GA, et al. Molecular Tumor Board: The University of California San Diego Moores Cancer Center Experience. Oncologist. 2014;19:631–6. DOI: 10.1634/theoncologist.2013-0405
[5] Khan O, Lim Choi Keung SN, Zhao L, Arvanitis TN. A Hybrid EAV-Relational Model for Consistent and Scalable Capture of Clinical Research Data. Integrating Information Technology and Management for Quality of Care. 2014:32–5. DOI: 10.3233/978-1-61499-423-7-32