Enhancing phenotype recognition in clinical notes using large language models: PhenoBCBERT and PhenoGPT

Jingye Yang; Cong Liu; Wendy Deng; Da Wu; Chunhua Weng; Yunyun Zhou; Kai Wang

doi:10.1016/j.patter.2023.100887

Enhancing phenotype recognition in clinical notes using large language models: PhenoBCBERT and PhenoGPT

Patterns (N Y). 2023 Dec 5;5(1):100887. doi: 10.1016/j.patter.2023.100887. eCollection 2024 Jan 12.

Authors

Jingye Yang^{1

2}, Cong Liu³, Wendy Deng¹, Da Wu¹, Chunhua Weng³, Yunyun Zhou^{1

4}, Kai Wang^{1

5}

Affiliations

¹ Raymond G. Perelman Center for Cellular and Molecular Therapeutics, Children's Hospital of Philadelphia, Philadelphia, PA 19104, USA.
² Department of Mathematics, University of Pennsylvania, Philadelphia, PA 19104, USA.
³ Department of Biomedical Informatics, Columbia University, New York, NY 10032, USA.
⁴ Biostatistics and Bioinformatics Facility, Fox Chase Cancer Center, Philadelphia, PA 19111, USA.
⁵ Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA 19104, USA.

Abstract

To enhance phenotype recognition in clinical notes of genetic diseases, we developed two models-PhenoBCBERT and PhenoGPT-for expanding the vocabularies of Human Phenotype Ontology (HPO) terms. While HPO offers a standardized vocabulary for phenotypes, existing tools often fail to capture the full scope of phenotypes due to limitations from traditional heuristic or rule-based approaches. Our models leverage large language models to automate the detection of phenotype terms, including those not in the current HPO. We compare these models with PhenoTagger, another HPO recognition tool, and found that our models identify a wider range of phenotype concepts, including previously uncharacterized ones. Our models also show strong performance in case studies on biomedical literature. We evaluate the strengths and weaknesses of BERT- and GPT-based models in aspects such as architecture and accuracy. Overall, our models enhance automated phenotype detection from clinical texts, improving downstream analyses on human diseases.

Keywords: BERT; GPT; Human Phenotype Ontology; clinical notes; electronic health records; named entity recognition; transformer.

Abstract

Grants and funding