Analysis of early screening results for high-risk populations of retinal diseases based on artificial Intelligence optical coherence tomography technology_Chinese Journal of Ocular Fundus Diseases

Authors：

Liu Dongning ¹ , Xu Xun ² , Zhao Junyi ¹ , Tian Na ¹ , An Yan ¹ , Liu Xinshu ¹ , Yu Jie ¹ , Yang Xue ¹ ,  Niu Tongtong ¹

1. Department of Ophthalmology, The Fourth People’s Hospital of Shenyang, Shenyang 110031, China;
2. Department of Ophthalmology, Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai 200080, China;

Corresponding?author：

Niu Tongtong, Email: 15840566565@163.com

Keywords：

Artificial intelligence; Screening; Wide-field; Fundus diseases; High-risk factors; Referral

DOI：

10.3760/cma.j.cn511434-20251009-00423

Video：

Export PDF Favorites Scan Get Citation

Abstract Full text Figures/Tables Video References Cited by

Objectives To evaluate and preliminarily analyze the application value and efficacy of artificial intelligence optical coherence tomography (AI-OCT) technology in the early screening of retinal diseases among the elderly, hypertension, hyperglycemia, high myopia and hyperlipidemia (referred to as "Five-High") population. Methods A diagnostic trial was conducted. A total of 3 834 patients (7 668 eyes) with "Five-High" risk factors who visited the outpatient clinics of Shenyang Fourth People’s Hospital from July to December 2024 were included. OCT imaging of the macular and peripheral retina was performed using the Bigway AI-OCT image analysis system (wide-field three-dimensional scanning mode). The deep learning-based system automatically identified and labeled eight types of high-risk retinal lesions: subretinal fluid (SRF), intraretinal fluid (IRF), epiretinal membrane (ERM), choroidal neovascularization (CNV), hyper-reflective foci (HRF), retinal pigment epithelium detachment, retinal hemorrhage, and macular hole (MH). The positive rate of AI screening and the distribution of high-risk lesions were analyzed. Consistency between AI-OCT screening results and ophthalmologist review was assessed using Cohen’s Kappa test. Logistic regression was used to identify independent predictors of positive AI-OCT screening. Referral and treatment rates were also analyzed. Results Among 3 834 cases involving 7 668 eyes, 803 cases (1 606 eyes) were positive in AI-OCT screening, with a positive rate of 20.9% (803/3 834), including 266 high-risk and 537 non-high-risk patients, respectively. The positive screening rates of patients with "five highs" were as follows: hyperlipidemia 25.2% (185/735), advanced age 24.9% (746/1 998), hyperglycemia 24.8% (345/1 392), hypertension 23.8% (228/956), and high myopia 19.0% (40/210). Among 1 606 positive eyes, 1 355 high-risk lesions were identified by consensus. Among them, ERM had the largest number of identifications (780, 57.6%), followed by HRF (255, 18.8%), and MH had the smallest number of identifications (7, 0.5%). Physicians randomly reexamined 1 352 cases and 2 704 eyes. The number of positive and negative eyes diagnosed was 753 and 1 952 respectively. The number of positive and negative eyes screened by AI-OCT was 828 and 1,876 respectively. There was an excellent consistency between AI-OCT screening and physician diagnosis (Kappa=0.866, P=0.011). Multivariate logistic regression analysis showed that age [odds ratio (OR) =1.071, P＜0.001], high myopia (OR=1.921, P=0.001), and hyperglycemia (OR=1.287, P=0.005) were independent predictors of positive AI-OCT screening. Among 1 355 high-risk lesions, a total of 703 were referred (referral rate 51.9%). The three lesions with the highest referral rates were SRF (71.1%, 27/38), IRF (69.2%, 54/78), and CNV (61.5%, 24/39), respectively. Among the 803 cases with positive AI-OCT screening, 385 cases (47.9%) actually received referral suggestions, 259 cases (32.3%) were eventually diagnosed, and 109 cases (13.6%) received treatment. Compared with low-risk patients, the referral rate and diagnosis rate of high-risk patients were significantly higher (χ²=6.87, 4.48; P＜0.05), but there was no statistically significant difference in the final treatment acceptance rate between groups (χ²=1.15, P=0.28). Conclusions The established AI-OCT based screening model for fundus diseases in the “Five-High” population effectively improves the detection rate of early-stage lesions and promotes a shift from universal to precision screening. Patients with positive screening results have obvious referral and treatment obstacles, which requires clinical attention.

1.	中華醫學會眼科學分會眼底病學組, 中國醫師協會眼科醫師分會眼底病專業委員會. 我國糖尿病視網膜病變臨床診療指南(2022年)——基于循證醫學修訂[J]. 中華眼底病雜志, 2023, 39(2): 99-124. DOI: 10.3760/cma.j.cn511434-20230110-00018.Fundus Disease Group of Ophthalmological Society of Chinese Medical Association Fundus Disease Group of Ophthalmologist Branch of Chinese Medical Doctor Association. Evidence-based guidelines for diagnosis and treatment of diabetic retinopathy in China (2022)[J]. Chin J Ocul Fundus Dis, 2023, 39(2): 99-124. DOI: 10.3760/cma.j.cn511434-20230110-00018.
2.	Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography[J]. Science, 1991, 254(5035): 1178-1181. DOI: 10.1126/science.1957169.
3.	De Fauw J, Ledsam JR, Romera-Paredes B, et al. Clinically applicable deep learning for diagnosis and referral in retinal disease[J]. Nat Med, 2018, 24(9): 1342-1350. DOI: 10.1038/s41591-018-0107-6.
4.	Schmidt-Erfurth U, Sadeghipour A, Gerendas BS, et al. Artificial intelligence in retina[J]. Prog Retin Eye Res, 2018, 67: 1-29. DOI: 10.1016/j.preteyeres.2018.07.004.
5.	Li Z, Guo C, Nie D, et al. Deep learning for automated glaucomatous optic neuropathy detection from ultra-widefield fundus images[J]. Br J Ophthalmol, 2021, 105(11): 1548-1554. DOI: 10.1136/bjophthalmol-2020-317327.
6.	Bellemo V, Lim ZW, Lim G, et al. Artificial intelligence using deep learning to screen for referable and vision-threatening diabetic retinopathy in Africa: a clinical validation study[J/OL]. Lancet Digit Health, 2019, 1(1): e35-e44[2019-05-02]. https://pubmed.ncbi.nlm.nih.gov/33323239/. DOI: 10.1016/S2589-7500(19)30004-4.
7.	Ting DSW, Cheung CY, Lim G, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes[J]. JAMA, 2017, 318(22): 2211-2223. DOI: 10.1001/jama.2017.18152.
8.	Li Z, Wang L, Wu X, et al. Artificial intelligence in ophthalmology: the path to the real-world clinic[J/OL]. Cell Rep Med, 2023, 4(7): 101095[2023-07-18]. https://pubmed.ncbi.nlm.nih.gov/37385253/. DOI: 10.1016/j.xcrm.2023.101095.
9.	Liu X, Zhao C, Wang L, et al. Evaluation of an OCT-AI-based telemedicine platform for retinal disease screening and referral in a primary care setting[J/OL]. Transl Vis Sci Technol, 2022, 11(3): 4[2022-03-02]. https://pubmed.ncbi.nlm.nih.gov/35254422/. DOI: 10.1167/tvst.11.3.4.
10.	Shekhawat NS, Niziol LM, Sharma SS, et al. The utility of routine fundus photography screening for posterior segment disease: a stepped-wedge, cluster-randomized trial in South India[J]. Ophthalmology, 2021, 128(7): 1060-1069. DOI: 10.1016/j.ophtha.2020.11.025.
11.	Wong TY, Sun J, Kawasaki R, et al. Guidelines on diabetic eye care: the International Council of Ophthalmology recommendations for screening, follow-up, referral, and treatment based on resource settings[J]. Ophthalmology, 2018, 125(10): 1608-1622. DOI: 10.1016/j.ophtha.2018.04.007.
12.	Vujosevic S, Aldington SJ, Silva P, et al. Screening for diabetic retinopathy: new perspectives and challenges[J]. Lancet Diabetes Endocrinol, 2020, 8(4): 337-347. DOI: 10.1016/S2213-8587(19)30411-5.
13.	Spaide RF, Curcio CA. Drusen characterization with multimodal imaging[J]. Retina, 2010, 30(9): 1441-1454. DOI: 10.1097/IAE.0b013e3181ee5ce8.
14.	Wu Z, Luu CD, Ayton LN, et al. Optical coherence tomography-defined changes preceding the development of drusen-associated atrophy in age-related macular degeneration[J]. Ophthalmology, 2014, 121(12): 2415-2422. DOI: 10.1016/j.ophtha.2014.06.034.
15.	Zhang Z, Deng C, Paulus YM. Advances in structural and functional retinal imaging and biomarkers for early detection of diabetic retinopathy[J/OL]. Biomedicines, 2024, 12(7): 1405[2024-06-25]. https://pubmed.ncbi.nlm.nih.gov/39061979/. DOI: 10.3390/biomedicines12071405.
16.	Schmidt-Erfurth U, Chong V, Loewenstein A, et al. Guidelines for the management of neovascular age-related macular degeneration by the European Society of Retina Specialists (EURETINA)[J]. Br J Ophthalmol, 2014, 98(9): 1144-1167. DOI: 10.1136/bjophthalmol-2014-305702.
17.	Flaxel CJ, Adelman RA, Bailey ST, et al. Age-related macular degeneration preferred practice pattern^?[J]. Ophthalmology, 2020, 127(1): 1-65. DOI: 10.1016/j.ophtha.2019.09.024.
18.	鄢聞嘉, 羅德倫, 馮加勁, 等. 眼底抗血管內皮生長因子藥物應用與創新[J]. 中華眼底病雜志, 2023, 39(8): 701-707. DOI: 10.3760/cma.j.cn511434-20220620-00371.Yan WJ, Luo DL, Feng JJ, et al. An update on anti-vascular endothelial growth factor therapy in retinal diseases[J]. Chin J Ocul Fundus Dis, 2023, 39(8): 701-707. DOI: 10.3760/cma.j.cn511434-20220620-00371.
19.	Wells JA, Glassman AR, Ayala AR, et al. Aflibercept, Bevacizumab, or Ranibizumab for diabetic macular edema: two-year results from a comparative effectiveness randomized clinical trial[J]. Ophthalmology, 2016, 123(6): 1351-1359. DOI: 10.1016/j.ophtha.2016.02.022.
20.	Gu C, Wang Y, Jiang Y, et al. Application of artificial intelligence system for screening multiple fundus diseases in Chinese primary healthcare settings: a real-world, multicentre and cross-sectional study of 4795 cases[J]. Br J Ophthalmol, 2024, 108(3): 424-431. DOI: 10.1136/bjo-2022-322940.
21.	Ruamviboonsuk P, Tiwari R, Sayres R, et al. Real-time diabetic retinopathy screening by deep learning in a multisite national screening programme: a prospective interventional cohort study[J/OL]. Lancet Digit Health, 2022, 4(4): e235-e244[2022-03-07]. https://pubmed.ncbi.nlm.nih.gov/35272972/. DOI: 10.1016/S2589-7500(22)00017-6.
22.	Xie Y, Nguyen QD, Hamzah H, et al. Artificial intelligence for teleophthalmology-based diabetic retinopathy screening in a national programme: an economic analysis modelling study[J/OL]. Lancet Digit Health, 2020, 2(5): e240-e249[2020-04-23]. https://pubmed.ncbi.nlm.nih.gov/33328056/. DOI: 10.1016/S2589-7500(20)30060-1.
23.	Cui T, Lin D, Yu S, et al. deep learning performance of ultra-widefield fundus imaging for screening retinal lesions in rural locales[J]. JAMA Ophthalmol, 2023, 141(11): 1045-1051. DOI: 10.1001/jamaophthalmol.2023.4650.
24.	Zhang J, Lin S, Cheng T, et al. RETFound-enhanced community-based fundus disease screening: real-world evidence and decision curve analysis[J/OL]. NPJ Digit Med, 2024, 7(1): 108[2024-04-30]. https://pubmed.ncbi.nlm.nih.gov/38693205/. DOI: 10.1038/s41746-024-01109-5.
25.	Shekhawat NS, Niziol LM, Sharma SS, et al. The utility of routine fundus photography screening for posterior segment disease: a stepped-wedge, cluster-randomized trial in South India[J]. Ophthalmology, 2021, 128(7): 1060-1069. DOI: 10.1016/j.ophtha.2020.11.025.

1. 中華醫學會眼科學分會眼底病學組, 中國醫師協會眼科醫師分會眼底病專業委員會. 我國糖尿病視網膜病變臨床診療指南(2022年)——基于循證醫學修訂[J]. 中華眼底病雜志, 2023, 39(2): 99-124. DOI: 10.3760/cma.j.cn511434-20230110-00018.Fundus Disease Group of Ophthalmological Society of Chinese Medical Association Fundus Disease Group of Ophthalmologist Branch of Chinese Medical Doctor Association. Evidence-based guidelines for diagnosis and treatment of diabetic retinopathy in China (2022)[J]. Chin J Ocul Fundus Dis, 2023, 39(2): 99-124. DOI: 10.3760/cma.j.cn511434-20230110-00018.
2. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomography[J]. Science, 1991, 254(5035): 1178-1181. DOI: 10.1126/science.1957169.
3. De Fauw J, Ledsam JR, Romera-Paredes B, et al. Clinically applicable deep learning for diagnosis and referral in retinal disease[J]. Nat Med, 2018, 24(9): 1342-1350. DOI: 10.1038/s41591-018-0107-6.
4. Schmidt-Erfurth U, Sadeghipour A, Gerendas BS, et al. Artificial intelligence in retina[J]. Prog Retin Eye Res, 2018, 67: 1-29. DOI: 10.1016/j.preteyeres.2018.07.004.
5. Li Z, Guo C, Nie D, et al. Deep learning for automated glaucomatous optic neuropathy detection from ultra-widefield fundus images[J]. Br J Ophthalmol, 2021, 105(11): 1548-1554. DOI: 10.1136/bjophthalmol-2020-317327.
6. Bellemo V, Lim ZW, Lim G, et al. Artificial intelligence using deep learning to screen for referable and vision-threatening diabetic retinopathy in Africa: a clinical validation study[J/OL]. Lancet Digit Health, 2019, 1(1): e35-e44[2019-05-02]. https://pubmed.ncbi.nlm.nih.gov/33323239/. DOI: 10.1016/S2589-7500(19)30004-4.
7. Ting DSW, Cheung CY, Lim G, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes[J]. JAMA, 2017, 318(22): 2211-2223. DOI: 10.1001/jama.2017.18152.
8. Li Z, Wang L, Wu X, et al. Artificial intelligence in ophthalmology: the path to the real-world clinic[J/OL]. Cell Rep Med, 2023, 4(7): 101095[2023-07-18]. https://pubmed.ncbi.nlm.nih.gov/37385253/. DOI: 10.1016/j.xcrm.2023.101095.
9. Liu X, Zhao C, Wang L, et al. Evaluation of an OCT-AI-based telemedicine platform for retinal disease screening and referral in a primary care setting[J/OL]. Transl Vis Sci Technol, 2022, 11(3): 4[2022-03-02]. https://pubmed.ncbi.nlm.nih.gov/35254422/. DOI: 10.1167/tvst.11.3.4.
10. Shekhawat NS, Niziol LM, Sharma SS, et al. The utility of routine fundus photography screening for posterior segment disease: a stepped-wedge, cluster-randomized trial in South India[J]. Ophthalmology, 2021, 128(7): 1060-1069. DOI: 10.1016/j.ophtha.2020.11.025.
11. Wong TY, Sun J, Kawasaki R, et al. Guidelines on diabetic eye care: the International Council of Ophthalmology recommendations for screening, follow-up, referral, and treatment based on resource settings[J]. Ophthalmology, 2018, 125(10): 1608-1622. DOI: 10.1016/j.ophtha.2018.04.007.
12. Vujosevic S, Aldington SJ, Silva P, et al. Screening for diabetic retinopathy: new perspectives and challenges[J]. Lancet Diabetes Endocrinol, 2020, 8(4): 337-347. DOI: 10.1016/S2213-8587(19)30411-5.
13. Spaide RF, Curcio CA. Drusen characterization with multimodal imaging[J]. Retina, 2010, 30(9): 1441-1454. DOI: 10.1097/IAE.0b013e3181ee5ce8.
14. Wu Z, Luu CD, Ayton LN, et al. Optical coherence tomography-defined changes preceding the development of drusen-associated atrophy in age-related macular degeneration[J]. Ophthalmology, 2014, 121(12): 2415-2422. DOI: 10.1016/j.ophtha.2014.06.034.
15. Zhang Z, Deng C, Paulus YM. Advances in structural and functional retinal imaging and biomarkers for early detection of diabetic retinopathy[J/OL]. Biomedicines, 2024, 12(7): 1405[2024-06-25]. https://pubmed.ncbi.nlm.nih.gov/39061979/. DOI: 10.3390/biomedicines12071405.
16. Schmidt-Erfurth U, Chong V, Loewenstein A, et al. Guidelines for the management of neovascular age-related macular degeneration by the European Society of Retina Specialists (EURETINA)[J]. Br J Ophthalmol, 2014, 98(9): 1144-1167. DOI: 10.1136/bjophthalmol-2014-305702.
17. Flaxel CJ, Adelman RA, Bailey ST, et al. Age-related macular degeneration preferred practice pattern^?[J]. Ophthalmology, 2020, 127(1): 1-65. DOI: 10.1016/j.ophtha.2019.09.024.
18. 鄢聞嘉, 羅德倫, 馮加勁, 等. 眼底抗血管內皮生長因子藥物應用與創新[J]. 中華眼底病雜志, 2023, 39(8): 701-707. DOI: 10.3760/cma.j.cn511434-20220620-00371.Yan WJ, Luo DL, Feng JJ, et al. An update on anti-vascular endothelial growth factor therapy in retinal diseases[J]. Chin J Ocul Fundus Dis, 2023, 39(8): 701-707. DOI: 10.3760/cma.j.cn511434-20220620-00371.
19. Wells JA, Glassman AR, Ayala AR, et al. Aflibercept, Bevacizumab, or Ranibizumab for diabetic macular edema: two-year results from a comparative effectiveness randomized clinical trial[J]. Ophthalmology, 2016, 123(6): 1351-1359. DOI: 10.1016/j.ophtha.2016.02.022.
20. Gu C, Wang Y, Jiang Y, et al. Application of artificial intelligence system for screening multiple fundus diseases in Chinese primary healthcare settings: a real-world, multicentre and cross-sectional study of 4795 cases[J]. Br J Ophthalmol, 2024, 108(3): 424-431. DOI: 10.1136/bjo-2022-322940.
21. Ruamviboonsuk P, Tiwari R, Sayres R, et al. Real-time diabetic retinopathy screening by deep learning in a multisite national screening programme: a prospective interventional cohort study[J/OL]. Lancet Digit Health, 2022, 4(4): e235-e244[2022-03-07]. https://pubmed.ncbi.nlm.nih.gov/35272972/. DOI: 10.1016/S2589-7500(22)00017-6.
22. Xie Y, Nguyen QD, Hamzah H, et al. Artificial intelligence for teleophthalmology-based diabetic retinopathy screening in a national programme: an economic analysis modelling study[J/OL]. Lancet Digit Health, 2020, 2(5): e240-e249[2020-04-23]. https://pubmed.ncbi.nlm.nih.gov/33328056/. DOI: 10.1016/S2589-7500(20)30060-1.
23. Cui T, Lin D, Yu S, et al. deep learning performance of ultra-widefield fundus imaging for screening retinal lesions in rural locales[J]. JAMA Ophthalmol, 2023, 141(11): 1045-1051. DOI: 10.1001/jamaophthalmol.2023.4650.
24. Zhang J, Lin S, Cheng T, et al. RETFound-enhanced community-based fundus disease screening: real-world evidence and decision curve analysis[J/OL]. NPJ Digit Med, 2024, 7(1): 108[2024-04-30]. https://pubmed.ncbi.nlm.nih.gov/38693205/. DOI: 10.1038/s41746-024-01109-5.
25. Shekhawat NS, Niziol LM, Sharma SS, et al. The utility of routine fundus photography screening for posterior segment disease: a stepped-wedge, cluster-randomized trial in South India[J]. Ophthalmology, 2021, 128(7): 1060-1069. DOI: 10.1016/j.ophtha.2020.11.025.

Chinese Journal of Ocular Fundus Diseases

Latest ArticlesAnalysis of early screening results for high-risk populations of retinal diseases based on artificial Intelligence optical coherence tomography technology

Abstract Full text Figures/Tables Video References Cited by

Format

Content