Property Graph for Exploring Similarities in Real-World Oncology Standards.

Area of Science:

• Health Informatics
• Oncology Data Standards

Background:

• Real-world oncology standards are mature, but documentation on their interrelationships is scarce.
• Standard similarities mapping is crucial for enabling data integration and sharing across diverse oncology data standards.

Purpose of the Study:

• To compare entities and standard values in real-world oncology standards using a property graph approach.
• To establish a baseline description of similarities and differences between oncology data standards.

Main Methods:

• Utilized a property graph model to represent and compare entities and values within oncology standards.
• Employed direct matching and automated text embedding similarity calculations to identify overlapping and complementary elements.
• Explored relationships within the property graph to enhance contextual understanding of standards.

Main Results:

• Empirical exploration confirmed that all analyzed oncology standards share core entities representing key domains like demographics, diagnostics, and therapeutics.
• Identified significant overlaps and complementary aspects between entities and values across different standards.
• Demonstrated the effectiveness of property graphs in modeling complex standards and facilitating relationship exploration.

Conclusions:

• The property graph approach effectively models complex oncology standards, enabling contextual understanding through relationship exploration.
• The findings demonstrate the feasibility of integrating diverse oncology data sources and achieving interoperable sharing of artifacts.
• Standard mapping is a foundational step for advancing data integration and sharing in oncology.