UTH-Olympia@BC8 Track 3: Adapting GPT-4 for Entity Extraction and Normalizing Responses to Detect Key Findings in Dysmorphology Physical Examination Observations

Abstract

In this paper, we present our approach for the Biocreative VIII Track 3: Genetic Phenotype Extraction from Dysmorphology Physical Examination Entries (genetic conditions in pediatric patients). The aim of this track is to extract and normalize key findings present in dysmorphology physical examinations. We report an automated system relying on OpenAI's most recent large language model (LLM): GPT-4 to retrieve named entities and their spans from observations and normalize retrieved entities to Human Phenotype Ontology (HPO) concepts using dictionary matching algorithms. Our reported systems achieved an F1 score of 0.82 for standard matching in subtask 3a and an F1 score of 0.72 for exact matching in subtask 3b.

This article is part of the Proceedings of the BioCreative VIII Challenge and Workshop: Curation and Evaluation in the era of Generative Models.