Table 5. The performance of the deep learning system vs. general ophthalmologists in detecting notable peripheral retinal lesions.

Item	Sensitivity (95% CI), %	Specificity (95% CI), %	Accuracy (95% CI), %
Ophthalmologist A	93.6 (89.6–97.6)	98.7 (97.8–99.6)	97.6 (96.5–98.7)
Ophthalmologist B	85.9 (80.0–91.8)	96.8 (95.4–98.2)	94.5 (92.8–96.2)
Deep learning system	98.7 (96.9–100)	99.2 (98.5–99.9)	99.1 (98.4–99.8)

CI, confidence intervals; A, the general ophthalmologist with 5-year working experience in physical examination center; B, the general ophthalmologist with 3-year working experience in physical examination center.