PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
May 10, 2026Scientific Reports2 citationsOpen Access

Automated Phenotyping of Ophthalmologic Diseases Using Small Language Models

Automated phenotyping of ophthalmologic diseases from routine medical records using small language models and the human phenotype ontology (HPO)

View Full Paper
Ask AI
Bookmark
Share

Authors

BTBinh Duong ThaiSASebastian ArensTRThomas Reinhard

Discussion

Loading...

Member takes

Overview

Randomized trial evaluates automated extraction of ophthalmic phenotypes, suggesting improved clinical data use.

Key Points

  • This research examines the effectiveness of the human phenotype ontology (HPO) in automated phenotyping within ophthalmology.
  • Developed an AI pipeline that integrates text segmentation and negation detection using a small language model (PHI-4).
  • Manually annotated 175 ophthalmic medical records with HPO terms to create a ground truth dataset.
  • Evaluated the AI pipeline's performance based on metrics like Jaccard similarity, precision, recall, and F1 score.
  • Identified a total of 342 HPO terms manually, with the AI pipeline retrieving 341 terms.
  • Achieved a median Jaccard similarity of 0.67, precision of 0.83, recall of 0.82, and F1 score of 0.80.
  • The AI pipeline effectively extracts standardized phenotypes, indicating potential for improved data management.

Cite This Study

Thai et al. (2026) studied this question.

synapsesocial.com/papers/6a0020cec8f74e3340f9b9b5https://doi.org/10.1038/s41598-026-51512-z
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Improving automated deep phenotyping through large language models using retrieval-augmented generation2025
  2. 2Automatic genetic phenotype normalization from dysmorphology physical examinations: an overview of the BioCreative VIII—Task 3 competition2025
  3. 3FastHPOCR: Pragmatic, fast and accurate concept recognition using the Human Phenotype Ontology2024 · 4 citations
  4. 4Harnessing generative AI to annotate the severity of all phenotypic abnormalities within the Human Phenotype Ontology2024 · 2 citations
  5. 5Population health management of human phenotype ontology2025