BACKGROUND: Artificial intelligence (AI)-based fundus imaging enables efficient large-scale retinal screening; however, its application in occupational populations remains limited. OBJECTIVE: To evaluate the utility of an AI-based fundus screening system in teachers and to investigate fundus tessellated density (FTD) as a potential imaging marker associated with myopia and occupational factors. METHODS: In this prospective observational study, 648 teachers from Beijing underwent fundus photography and systemic examinations. AI algorithms were used to quantify FTD and other retinal parameters. Associations between spherical equivalent (SE), FTD, and systemic and occupational variables were assessed using univariate and multivariable regression analyses. Receiver operating characteristic (ROC) analysis was performed to evaluate the discriminative ability of FTD for high myopia. RESULTS: The mean SE was - 2.43 ± 2.94 D, and mean FTD was 0.0609 ± 0.0743. FTD showed a non-uniform distribution across the macula, with higher values in the nasal quadrant. Multivariable analysis demonstrated that FTD, near-work duration at 30 cm, and ocular structural parameters were independently associated with SE (all P 1 h) at 30 cm was associated with increased FTD. ROC analysis indicated moderate discriminative performance for high myopia (AUC = 0.7788 for overall FTD). In addition, FTD was associated with systemic immune-inflammatory indices, including NLR and PLR. CONCLUSIONS: AI-derived FTD may serve as a potential imaging marker associated with high myopia and occupational factors. These findings are observational and hypothesis-generating, and further longitudinal studies are required for validation.
Xiao et al. (2026) studied this question.