Table 1

Summary table for the different DL systems in the detection of referable diabetic retinopathy, glaucoma suspect, age-related macular degeneration and retinopathy of prematurity using fundus photographs

DL systemsYearTest data setsTest images (n)CNNAUCSensitivity (%)Specificity (%)
Referable diabetic retinopathy
 Abràmoff et al 14 2016Messidor-21748AlexNet/VGG0.9896.8087.00
 Gulshan et al 12 2016Messidor-21748Inception-V30.998798.50
96.1093.90
EyePACS-199630.99190.3098.10
97.5093.40
 Gargeya and Leng15 2017Kaggle images75 137Customised CNN0.97NANA
E-Ophtha4630.96NANA
Messidor-217480.94NANA
 Ting et al 11 2017SiDRP 14–1571 896VGG-190.93690.5091.60
Guangdong15 7980.94998.7081.60
SIMES30520.88997.1082.00
SINDI45120.91799.373.3
SCES19360.91910076.30
BES10520.92994.4088.50
AFEDS19680.9898.8086.50
RVEEH23020.98398.9092.20
Mexican11720.9591.8084.80
CUHK12540.94899.383.10
HKU77060.96410081.30
 Abràmoff et al 28 201810 primary care practice sites from the USA892 patientsAlex/VGGNA87.290.7
Glaucoma suspect*
 Ting et al 11 2017SiDRP 14–1571 896VGG-190.94296.4093.20
 Li et al 16 2018Guangdong48 1160.98695.6092.00
Age-related macular degeneration
 Ting et al 11 2017SiDRP 14–1535 948VGG-190.93293.2088.70
 Burlina et al 17 2017AREDS120 656AlexNet, OverFeat0.940–0.96NANA
 Grassmann et al 18 2018AREDS120 656AlexNet, GoogleNet, VGG, Inception-V3, ResNet, Inception-ResNet-V2NA84.2094.30
Retinopathy of prematurity
 Brown et al 19 2018i-ROP100Inception-V1 and U-NetNA10094
  • The diagnostic performance is not comparable between the different DL systems given the different data sets used in the individual study.

  • *Definition of glaucoma suspect: (1) Ting et al 11—vertical cup to disc ratio of 0.8 or greater, and any glaucomatous disc changes; (2) Li et al 16—vertical cup to disc ratio of 0.7 or greater, and any glaucomatous disc changes.

  • AFEDS, African American Eye Disease Study; AREDS, Age-Related Eye Disease Study; AUC, area under the receiver operating characteristic curve; BES, Beijing Eye Study; CNN, convolutional neural network; CUHK, Chinese University Hong Kong; DL, deep learning; SiDRP 14–15, Singapore Integrated Diabetic Retinopathy Screening Programme; HKU, Hong Kong University; NA, not available; RVEEH, Royal Victorian Eye and Ear Hospital; SCES, Singapore Chinese Eye Study; SIMES, Singapore Malay Eye Study; SINDI, Singapore Indian Eye Study.