EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Benchmarking diffusion models against state-of-the-art architectures for OCT fluid biomarker segmentation

Katherine Du, Utkarsh Doshi, Benjamin DiCenzo, Jessica Jiang, Ethan Wu, Adarsh Gadari, Sharat Chandra Vupparaboina, Elham Sadeghi, Sandeep Chandra Bollepalli, José-Alain Sahel, Jay Chhablani, Kiran Kumar Vupparaboina iD

PLOS One · 2025

DOI: 10.1371/journal.pone.0335615 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, 2025. Verfasst von Katherine Du und 11 weiteren Autor:innen. Beteiligte Einrichtung: Department of Ophthalmology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, United States of America. Hauptthemen laut MEDLINE: Tomography, Optical Coherence, Subretinal Fluid, Retinal Diseases. Die Arbeit nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NEI NIH HHS, Eye and Ear Foundation of Pittsburgh, Research to Prevent Blindness.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objectives

Retinal diseases, major causes of vision impairment and blindness, are assessed using optical coherence tomography (OCT) scans. Automated report generation for retinal OCT scans, powered by deep learning, can help standardize interpretations and track retinal disease in clinical settings. A key challenge is accurately segmenting retinal disease signatures. This study explores using the diffusion model to segment subretinal fluid (SRF), intraretinal fluid (IRF), and pigment epithelial detachment (PED) in typical clinical settings, comparing their performance to other leading segmentation models.

Methods

We labeled OCT scans and extracted those with specific pathologic retinal features: 269 scans with SRF, 224 scans with IRF, and 114 scans with PED. Three trained reviewers manually segmented these features for downstream analysis. Using manually segmented scans as the ground truth, we trained the diffusion model, Nested U-Net, nnU-Net, TransUNet, and SwinUNet to predict these segmentations. All models were evaluated using 5-fold cross-validation, with performance measured by Dice coefficient, sensitivity, specificity, Pearson correlation coefficient, and R 2 .

Results

All models show high similarly with ground truth segmentations in predicting SRF, IRF, and PED, as shown by the Dice coefficient (Diffusion model: 0.81 ± 0.12, 0.66 ± 0.09, 0.75 ± 0.11). The diffusion model has relatively higher sensitivity compared to most other models, while all models display very high specificity. The Pearson correlation coefficient and R 2 values show strongly associated pixel quantification of segmented areas for models, with the nnU-Net model performing the strongest overall.

Conclusion

This study demonstrates that while diffusion models can comparably segment retinal pathologies using a limited number of manually annotated scans, the nnU-Net model remains the most effective overall for automated OCT analysis.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Tomography, Optical CoherenceSubretinal FluidRetinal DiseasesHumansBiomarkersBenchmarkingDeep LearningRetinaRetinal Detachment

Substanzen

Biomarkers

Autor:innen und Institutionen

  • Department of Ophthalmology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, United States of America.

ORCID-Profile: Kiran Kumar Vupparaboina

Förderung

  • NEI NIH HHS
  • Eye and Ear Foundation of Pittsburgh
  • Research to Prevent Blindness
  • NIH
  • The Hillman Challenge Grant - Exploratory Research Project

Zitieren

APA

Du, K., Doshi, U., DiCenzo, B., Jiang, J., Wu, E., Gadari, A., Vupparaboina, S. C., Sadeghi, E., Bollepalli, S. C., Sahel, J. A., Chhablani, J., & Vupparaboina, K. K. (2025). Benchmarking diffusion models against state-of-the-art architectures for OCT fluid biomarker segmentation. PLOS One. https://doi.org/10.1371/journal.pone.0335615

Vancouver

Du K, Doshi U, DiCenzo B, Jiang J, Wu E, Gadari A, et al. Benchmarking diffusion models against state-of-the-art architectures for OCT fluid biomarker segmentation. PLOS One. 2025. doi:10.1371/journal.pone.0335615

BibTeX
@article{Du2025,
  author = {Du, Katherine and Doshi, Utkarsh and DiCenzo, Benjamin and Jiang, Jessica and Wu, Ethan and Gadari, Adarsh and Vupparaboina, Sharat Chandra and Sadeghi, Elham and Bollepalli, Sandeep Chandra and Sahel, José-Alain and Chhablani, Jay and Vupparaboina, Kiran Kumar},
  title = {Benchmarking diffusion models against state-of-the-art architectures for OCT fluid biomarker segmentation},
  journal = {PLOS One},
  year = {2025},
  doi = {10.1371/journal.pone.0335615}
}
Kostenlos starten →