Knowledge Representation of a Multicenter Adolescent and Young Adult Cancer Infrastructure: Development of the STRONG...

Area of Science:

• Health Informatics
• Bioinformatics
• Data Science

Background:

• Rare disease research, particularly for adolescent and young adults (AYAs) with cancer, requires large, dispersed data collection efforts.
• Data harmonization presents significant challenges, often rendering data incompatible and hindering research progress.
• The STRONG Adolescent and Young Adult (AYA) project aims to establish a federated data infrastructure for AYA cancer research, facing diverse data formats.

Purpose of the Study:

• To demonstrate the application of health care-agnostic semantic web technologies to overcome data interoperability and harmonization challenges.
• To detail the methods used within the STRONG AYA project to create a federated data infrastructure.
• To showcase how semantic technologies can enable the integration of varied data formats in rare disease research.

Main Methods:

• Knowledge graphs were developed to structure case-mix and core outcome measures for the STRONG AYA project.
• A semantic map was created using knowledge graphs and an annotation helper plugin for the Flyover tool.
• Flyover, a tool for converting data into Resource Description Framework (RDF) triples, was utilized to enable semantic interoperability.

Main Results:

• Knowledge graphs provided a comprehensive overview of the numerous concepts within the STRONG AYA project.
• Semantic terminology mapping and an annotation helper facilitated querying of data with complex terminologies without data modification.
• Knowledge graphs and the semantic map were published on a Hugo webpage to enhance transparency and understanding.

Conclusions:

• Semantic web technologies, including RDF and knowledge graphs, offer a flexible solution for data interoperability and reusability in federated cancer data infrastructures.
• These domain-agnostic technologies successfully linked semantically meaningful concepts to nonstandardized healthcare data, making it interoperable.
• The approach overcomes limitations of rigid standardized schemas, proving viable for complex research initiatives like the STRONG AYA project.