Purpose:To assess the accuracy and robustness of the AI algorithm for detecting referable diabetic retinopathy (RDR), referable macular diseases (RMD), and glaucoma suspect (GCS) from fundus images in community and in-hospital screening scenarios.MethodsWe collected two color fundus image datasets, namely, PUMCH (556 images, 166 subjects, and four camera models) and NSDE (534 images, 134 subjects, and two camera models). The AI algorithm generates the screening report after taking fundus images. The images were labeled as RDR, RMD, GCS, or none of the three by 3 licensed ophthalmologists. The resulting labels were treated as “ground truth” and then were used to compare against the AI screening reports to validate the sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) of the AI algorithm.ResultsOn the PUMCH dataset, regarding the prediction of RDR, the AI algorithm achieved overall results of 0.950 ± 0.058, 0.963 ± 0.024, and 0.954 ± 0.049 on sensitivity, specificity, and AUC, respectively. For RMD, the overall results are 0.919 ± 0.073, 0.929 ± 0.039, and 0.974 ± 0.009. For GCS, the overall results are 0.950 ± 0.059, 0.946 ± 0.016, and 0.976 ± 0.025.ConclusionThe AI algorithm can work robustly with various fundus camera models and achieve high accuracies for detecting RDR, RMD, and GCS.

Purpose Evaluate the efficiency of using an artificial intelligence reading label system in the diabetic retinopathy grading training of junior ophthalmology resident doctors and medical students. Methods Loading 520 diabetic retinopathy patients’ colour fundus images into the artificial intelligence reading label system. Thirteen participants, including six junior ophthalmology residents and seven medical students, read the images randomly for eight rounds. They evaluated the grading of images and labeled the typical lesions. The sensitivity, specificity, and kappa scores were determined by comparison with the participants’ results and diagnosis gold standards. Results Through eight rounds of reading, the average kappa score was elevated from 0.67 to 0.81. The average kappa score for rounds 1 to 4 was 0.77, and the average kappa score for rounds 5 to 8 was 0.81. The participants were divided into two groups. The participants in Group 1 were junior ophthalmology resident doctors, and the participants in Group 2 were medical students. The average kappa score of Group 1 was elevated from 0.71 to 0.76. The average kappa score of Group 2 was elevated from 0.63 to 0.84. Conclusion The artificial intelligence reading label system is a valuable tool for training resident doctors and medical students in performing diabetic retinopathy grading.

Purpose To assess the association of genes in the high-density lipoprotein metabolic pathway (HDLMP) with polypoidal choroidal vasculopathy (PCV) and the genetic difference in the HDLMP between PCV and age-related macular degeneration (AMD). Methods We performed a literature search in EMBASE, PubMed, and Web of Science for genetic studies on 7 single nucleotide polymorphisms (SNPs) from 5 genes in the HDLMP including cholesteryl ester transfer protein (CETP), hepatic lipase (LIPC), lipoprotein lipase (LPL), ATP-binding cassette transporter A1 (ABCA1), and ATP-binding cassette transporter G1 (ABCG1) in PCV. All studies were published before September 30, 2017, without language restriction. Pooled odds ratios (ORs) and 95% confidence intervals (CIs) of each polymorphism were estimated. We also compared the association profiles between PCV and AMD and performed a sensitivity analysis. Results Our result is based on 43 articles. After excluding duplicates and articles without complete information, 7 studies were applicable to meta-analysis. 7 polymorphisms were meta-analyzed: CETP rs2303790/rs3764261, LIPC rs10468017/rs493258, LPL rs12678919, ABCA1 rs1883025, and ABCG1 rs57137919. We found that in Asian population, CETP rs3764261 (T allele; OR = 1.46; 95% CI: 1.28–1.665, P < 0.01), CETP rs2303790 (G allele; OR = 1.57; 95% CI: 1.258–1.96, P < 0.01), and ABCG1 rs57137919 (A allele; OR = 1.168; 95% CI: 1.016–1.343, P < 0.01) were significantly associated with PCV, and ABCG1 rs57137919 (A allele; OR = 1.208, 95% CI: 1.035–1.411, P < 0.01) has different effects in PCV and AMD. The other 4 polymorphisms in LIPC/LPL/ABCA1 had no significant association with PCV (P > 0.05). The sensitivity analysis validated the significance of our analysis. Conclusions Our study revealed 7 polymorphisms in 5 genes. Among them, CETP (rs3764261/rs2303790) and ABCG1 (rs57137919) were the major susceptibility genes for PCV in Asian population and ABCG1 (rs57137919) showed allelic diversity between PCV and AMD. Since the size for PCV and AMD was small, we need to study these genes genotyping in larger samples.