OntologyAligner improves mapping of biomedical text to concepts
OntologyAligner: Ontology-Aligned Retrieval and Hierarchy-Guided Large Language Model Reranking for Biomedical Ontology Normalization
Artificial IntelligenceComputation and Language
Summary
Mapping medical terms from free text to standardized concepts can be tricky because words vary and related concepts are often similar. The authors created OntologyAligner, a three-step method that finds potential matches, ranks them using large language models, and then refines choices by considering concept hierarchies. They tested their method on many examples related to human health traits and showed it works better than previous approaches. The method also works well on other biomedical concept databases, making it broadly useful.
What this means in practice
- •For hospital data teams: Improve electronic health record annotation by accurately linking patient text to standard biomedical concepts for better data analysis.
- •For healthcare software developers: Build clinical decision support tools that rely on precise normalization of biomedical terms to structured ontologies using OntologyAligner’s framework.
Authors
Jie Song, Zhichuan Xu, Ziyu Lu, Meng Xiao, Cheng Bi, Yuxin Zhang, Xin Zheng, Xiaoran Li, Qiongfang Cao, Hao Yang, Bairong Shen
Abstract
Biomedical ontology normalization maps free-text expressions to standardized concepts, enabling consistent integration and analysis of biomedical data. This task remains challenging because lexical variation and subtle distinctions among hierarchically related concepts can obscure concept boundaries. We present OntologyAligner, a three-stage framework that combines ontology-aligned retrieval, large language model candidate reranking, and selective hierarchy-guided refinement. We also construct PhenoNormBench, a unified benchmark comprising 13,390 samples from seven Human Phenotype Ontology datasets. OntologyAligner achieved state-of-the-art performance on HPO normalization, with 88.78% Macro Top-1 Accuracy and 86.75% Micro Top-1 Accuracy, exceeding the strongest baseline by 4.85 and 5.07 percentage points, respectively. Ablation analyses showed complementary contributions from all three stages, and sensitivity analyses demonstrated stability across candidate-set sizes and model backbones. Applications to MONDO, MEDIC, and NCBITaxon further established portability to other ontologies. OntologyAligner offers a generalizable framework for accurate mapping of biomedical text to structured ontology concepts. PhenoNormBench and the code are publicly available at https://github.com/zhelishisongjie/OntologyAligner.