Matters arising: Utilizing foundation models for developing clinical tools.
editorial · Level V
Where this comes from
- Record sourced from PubMed, PMID 41254183.
- Also identified by DOI 10.1038/s41746-025-02065-4 and PMC identifier 12627685.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
Zhang et al. developed a RETFound-enhanced deep learning model to detect multiple eye diseases and tested it in a community-based screening setting. Nevertheless, we believe that more information could be provided to support the claim that the proposed model has better generalizability compared to conventional convolutional neural networks adopted by two commercial models. Thus, we raised four points regarding more details of other models, generalizability, the fine-tuning dataset, and statistical analysis.