Construction and application of Chinese breast cancer knowledge graph based on multi-source heterogeneous data

Area of Science:

• Medical Informatics
• Oncology
• Artificial Intelligence

Background:

• General medical knowledge graphs are vast but challenging to manually curate, limiting their use in medical intelligence.
• Breast cancer diagnosis and treatment require efficient AI-driven solutions to improve patient outcomes.

Purpose of the Study:

• To develop a framework for constructing a specialized breast cancer knowledge graph (BCKG).
• To integrate heterogeneous data sources for comprehensive breast cancer information.

Main Methods:

• Extracting knowledge triples from clinical guidelines, medical encyclopedias, and electronic medical records.
• Fusing extracted triples from various sources to build the BCKG.

Main Results:

• The constructed BCKG supports knowledge-based question answering for breast cancer.
• BCKG facilitates improved breast cancer postoperative follow-up and healthcare management.
• Demonstrated enhancement in the quality and efficiency of breast cancer diagnosis and treatment.

Conclusions:

• The proposed framework effectively builds a breast cancer knowledge graph from heterogeneous data.
• BCKG shows potential for advancing AI applications in breast cancer care, from diagnosis to long-term management.