Performance of Artificial Intelligence in Detecting Diabetic Macular Edema From Fundus Photography and Optical Coherence Tomography Images: A Systematic Review and Meta-analysis
Authors' conclusion
This meta-analysis demonstrates satisfactory performance of AI in detecting DME from FP or OCT images. External validation is warranted for future studies to evaluate model generalizability. Further investigations may estimate optimal sample size, effect of class balance, patient demographics, and additional benefits of OCT volumetric scans.
Study context
- Population
- Studies using AI to detect diabetic macular edema (DME) from fundus photography (FP) or optical coherence tomography (OCT) images; 7 electronic libraries searched up to 12 February 202353 studies
- Comparison
- AI on FP vs AI on OCT images for detecting DME
- FP-based algorithmsn = 25 studies
- OCT-based algorithmsn = 28 studies
- Outcome
- Performance of AI in detecting DME (pooled AUROC, sensitivity, specificity)
Methodological detailsAnalyses limited by unstandardized algorithm outcomes and insufficient data on patient demographics, OCT volumetric scans and external validation
- Analyses limited by unstandardized algorithm outcomes and insufficient data on patient demographics, OCT volumetric scans and external validation
Pooled result
As reported in the abstract
Pooled diagnostic performance for DME detection
| Strategy | AUROC | Sensitivity | Specificity |
|---|---|---|---|
| FP-based algorithms | 0.964 | 92.6% | 91.1% |
| OCT-based algorithms | 0.985 | 95.9% | 97.9% |
AUROC
FP-based algorithms0.964
OCT-based algorithms0.985
Sensitivity
FP-based algorithms92.6%
OCT-based algorithms95.9%
Specificity
FP-based algorithms91.1%
OCT-based algorithms97.9%
Further pooled results
Factors affecting model performance
- Potential factors improving performance: deep learning techniques, larger size and more diversity in training data sets
- Better performance on internal than external validation; models trained with multiple data sets showed better results on external validation
Original publication
Performance of Artificial Intelligence in Detecting Diabetic Macular Edema From Fundus Photography and Optical Coherence Tomography Images: A Systematic Review and Meta-analysis. Diabetes Care. Published 19 January 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →