Table 3.

Performance in papilledema detection and grading

Author (year)	AUC (95% CI)	Sensitivity (%)	Specificity (%)	Accuracy (%)	Secondary results
Papilledema detection
Fatima et al.[12] (2017)	-	83.94	88.39	85.89
Akbar et al.[13] (2017)	-	90.01	96.39	92.86
Milea D et al.[9] (2020)	0.96 (0.95–0.97)	96.4	84.7	87.5
Biousse V et al.[10] (2020)	0.96 (0.94–0.97)	83.1 not statistically different from Expert 1and 2	94.3 significantly better than expert 1 (P<0.001), identical to Expert 2	91.5 not statistically different from Expert 1and 2	Time needed: DLS: 25 s Expert 1: 61 min Expert 2: 74 min
Saba T et al.[14] (2021)	-	98.63	97.83	99.17
Papilledema grading
Akbar S et al.[13] (2017)	-	97.32	96.90	97.85
Saba T et al.[14] (2021)	-	99.82	98.65	99.89
Vasseneix C et al.[11] (2021)	0.93 (0.89–0.96)	91.8 comparable to neuro-ophthalmologists 91.8 (P=1)	82.6 comparable to neuro-ophthalmologists 73.9 (P=0.09)	87.9 comparable to neuro-ophthalmologists 84.1 (P=0.19)

AUC: Area under the curve, CI=Confidence interval, DLS=Deep learning systems