Table 6. Performance of the 3 ensembling binary classifiers at the B-scan level.
CNV vs NORMAL | DME vs NORMAL | DRUSEN vs NORMAL | |
---|---|---|---|
Accuracy (95% CI) | 0.983 (0.981–0.985) | 0.988 (0.987–0.989) | 0.978 (0.976–0.980) |
Sensitivity (95% CI) | 0.984 (0.981–0.987) | 0.986 (0.984–0.988) | 0.981 (0.980–0.982) |
Specificity (95% CI) | 0.993 (0.992–0.994) | 0.991 (0.990–0.992) | 0.986 (0.984–0.998) |
AUC (95% CI) | 0.997 (0.996–0.998) | 0.996 (0.995–0.997) | 0.993 (0.991–0.995) |
CI: confidence interval; CNV: choroidal neovascularization; DME: diabetic macular edema; AUC: area under receiver operating characteristic curve.