Skip to main content

Table 7 Comparison of the performance of referable DR grading from different systems in clinic settings

From: Implementing and evaluating a fully functional AI-enabled model for chronic eye disease screening in a real clinical environment

IDx-DR

(DR)

EyeArt

(DR)

SELENA+

(DR)

Offline AI System*

(DR)

VoxelCloud Retina

(DR)

RAIDS

(DR, AMD, and Glaucoma)

TeleEye MD (DR, AMD, and Glaucoma)

Validated on 900 patients

SE: 87.2%;

SP: 90.7%

IDx-DR 2.0

Validated on 1415 patients in the Hoorn Diabetes Care System

SE: 91%

SP: 84%

(Using EURODIAB criteria)

EyeArt v2.1

Validated on 30,405 patients in the English Diabetic Eye Screening Programme in the UK

SE: 95.7%

SP: 54.0%

Validated on 893 patients

SE: 95.5%

SP: 85.3%

Validated on 1574 patients

in a mobile screening program in Zambia

SE: 92.3%

SP: 89.0%

Validated on 69 patients in Australia

SE: 96.9%

SP: 87.7%

Validated on 236 participants in Australia

SE: 96.9%

SP: 87.7%

Validated on 15,805 patients at 155 diabetes centres in China

SE: 83.3%

SP: 92.5%

Validated on 110,784 participants from 65 healthcare centers in China

DR

SE: 83.7%

SP: 98.6%

AMD

SE: 88.1%

SP: 97.6%

Referral possible glaucoma

SE: 81.3%

SP: 95.0%

Validated on 231

Patient in this study

DR

SE: 100.0%

SP: 80.1%

AMD

SE: 90.9%

SP: 78.8%

Referral possible glaucoma

SE: 93.2%

SP: 76.8%

  1. * From the Centre for Eye Research, Australia