Table 1 Summary of the data obtained from the included studies
Author | Year | Country | Sample size (patient) | Eye disease | Reference standard (RS) | SN and SP of (RS) | OCT or funds image | AI system | Country of AI system | SN and SP of AI |
|---|---|---|---|---|---|---|---|---|---|---|
Lim JI et al. [6] | 2022 | US | 521 | DR and ME | General ophthalmologists and retina specialists | Retina specialists: SN 59.5%, SP 98.9% General ophthalmologists: SN 20.6%, SP 99.8% | fundus image | EyeArt | US | SN: 96.4%, SP: 88.4% |
Zhang Wf et al. [7] | 2022 | China | 630 | DR and ME | Ophthalmologists | N/A | Funds image | EyeWisdom V1 | China | SN: 98.23% SP: 74.45 |
Liu R et al. [8] | 2022 | China | 600 | DR and ME | Retina specialists | N/A | OCT and fundus image | Faster R-CNN | China | ? |
Li N et al. [9] | 2021 | China | 1147 | DR | Retina specialist | N/A | Funds image | VoxelCloud | China | SN:85.1% SP:95.6% |
Dow ER et al. [10] | 2023 | US | 1222 | DR | Retina specialists | SN: 69.5% SP: 96.9% | Funds image | IDx-DR | Us | SN: 95.5% SP: 60.3% |
Vebraak Fd et al. [11] | 2019 | ? | 1425 | DR | Retina specialists | N/A | Funds image | IDx-DR-EU-2.1 | The Netherlands | SN: 79.4% SP: 93.8% |
Wintergerst Mwm et al. [12] | 2021 | Germany | 75 | DR and ME | Trained graders and ophthalmologist experts | SN: 67% SP:74 | Funds image | EyeArt version 2.1. | US | SN: 100% SP: 98 |
Abràmoff Md et al. [13] | 2018 | US | 819 | DR and ME | Trained grader | N/A | OCT and fundus | IDx-D | US | SN: 87.2% SP: 90.7 |
Quellec [14] | 2019 | France | 164,660 (procedures) | DR | Certified ophthalmologists | N/A | Funds image | OphthAI | France | SN:99% SP:90.2 |
Wang Y et al. [15] | 2021 | China | 4900 | DR | Trained graders and licensed and senior ophthalmologists | N/A | Funds image | DLA based on inceotion -V3 convolutional neural networks | China | SN: 97% and SP: 87.9% |
Baget-Bernaldiz M et al. [16] | 2021 | Spain | 7164 | DR | Retina specialists | N/A | Funds image | DLA built on convolutional neural networks. | Spain | SN: 97.92% SP:99.91 |
Nunez do Rio JM et al. [17] | 2022 | India | 11,199 | DR | Trained optometrists and ophthalmologists | N/A | Funds image | Zeiss VISUHEALTH-AI DR | Singapore | SN: 72.08% SP: 85.65 |
Gulshan V et al. [18] | 2019 | India | 3049 | DR and ME | Trained graders and retinal specialist | Retinal Specialist (Aravind): SN: 89.8% Sp: 83.5% Retinal Specialist (Sankara): SN: 73.4% SP: 98.7% Trained Grader (Aravind): SN: 75.7% Sp: 94.2% Trained Grader (Sankara): SN: 84.2% Sp: 98.6 | Funds image | Automated DR grading system based on deep learning algorithms | USA | ? |
Rogers T.W et al. [19] | 2020 | Mexico | 5752 | DR | Board od experts | N/A | Funds image | Pegasus, Visulytix Ltd., U | UK | SN:81.6% SP:81.7 |
Scheetz J et al. [20] | 2021 | China | 203 | DR | NHS-certified retinal grader | N/A | Funds image | ? | Australia | SN:96.9% SP:87.7 |
Zhang Y et al. [21] | 2020 | China | 4726 | DR and ME | Ophthalmologis | N/A | Funds image | VoxelCloud | China | SN: 83.3%. SP: 92.5 |
Wongchaisuwat et al. [22] | 2021 | Thailand | Nidek: 3515 Eidon:2663 | DR and ME | Retinal experts | N/A | Funds image by Nidek and Eidon cameras | DL based on convolutional neural networks | Thailand | SN of AI: Nidek: 93% Eidon: 88% SP of AI: Nidek: 91% Eidon: 85 |
Ming s et al. [23] | 2021 | China | 193 | DR | Retinal experts | N/A | Funds image | China | Eye wisdom | SN: 83.3% SP: 97.9% |