Advanced and interpretable corneal staining assessment through fine grained knowledge distillation
The assessment of corneal fluorescein staining is essential, yet current AI models for Corneal Staining Score (CSS) assessments inadequately identify punctate lesions due to annotation challenges and noise, risk misrepresenting treatment responses through "plateau" effects, and highlight the necessity for real-world evaluations to enhance disease severity assessments. To address these limitations, we developed the Fine-grained Knowledge Distillation Corneal Staining Score (FKD-CSS) model. FKD-CSS integrates fine-grained features into CSS grading, providing continuous and nuanced scores with interpretability. Trained on corneal staining images collected from dry eye (DE) patients across 14 hospitals, FKD-CSS achieved robust accuracy, with a Pearson's r of 0.898 and an AUC of 0.881 in internal validation, matching senior ophthalmologists' performance. External tests on 2376 images from 23 hospitals across China further validated its efficacy (r: 0.844-0.899, AUC: 0.804-0.883). Additionally, FKD-CSS demonstrated generalizability in multi-ocular-surface-disease testing, underscoring its potential in handling different staining patterns.
Authors
- Pujin Cheng (ORCID: https://orcid.org/0009-0000-4844-5146)
- Yu Kang (ORCID: https://orcid.org/0000-0002-8706-3252)
- Tingting Zhang (ORCID: https://orcid.org/0000-0002-1919-1471)
- Hongliang Xue (ORCID: https://orcid.org/0000-0002-9437-6396)
- Xiaoyi Li (ORCID: https://orcid.org/0000-0002-1453-4359)
- Yansong Huang
- Yan Lou
- Junyan Lyu (ORCID: https://orcid.org/0000-0001-5744-5201)
- Xiaoying Tang (ORCID: https://orcid.org/0000-0002-7549-6560)
- Zhonghua Wang (ORCID: https://orcid.org/0000-0001-6301-9698)
- Yuqing Deng
- Xiaoqing Hu (ORCID: https://orcid.org/0000-0001-5007-4567)
- Jin Yuan (ORCID: https://orcid.org/0000-0001-5756-5414)
- Yimin Zhang (ORCID: https://orcid.org/0000-0003-0724-5142)
- Shi-you Zhou
- Kenneth K. Y. Wong (ORCID: https://orcid.org/0000-0003-3668-3539)
- Ruiwen Xu
- Lirong Ling
Institutions
- Sun Yat-sen University (CN)
- The University of Queensland (AU)
- Guangdong Pharmaceutical University (CN)
- Southern University of Science and Technology (CN)
- Jiaxing University (CN)
- Key Laboratory of Guangdong Province (CN)
- First Affiliated Hospital of Guangzhou Medical University (CN)
- University of Hong Kong (HK)
- Guangzhou Medical University (CN)
Publication Details
- Journal
- npj Digital Medicine
- Published
- 2025-05-23
- DOI
- https://doi.org/10.1038/s41746-025-01706-y
- Citations
- 2
- Primary Topic
- Ocular Surface and Contact Lens
- Type
- article
- Field-Weighted Citation Impact
- 3.76
Funders
- National Natural Science Foundation of China
- Sun Yat-sen University
- China Postdoctoral Science Foundation
- Shenzhen Municipal Science and Technology Innovation Council
- National Key Research and Development Program of China
- Fundamental Research Funds for the Central Universities