27 February 2025 in npj Digital Medicine
Glaucoma, a leading cause of blindness, requires accurate early detection. We present an AI-based Glaucoma Screening (AI-GS) network comprising six lightweight deep learning models (total size: 110 MB) that analyze fundus images to identify early structural signs such as optic disc cupping, hemorrhages, and nerve fiber layer defects. The segmentation of the optic cup and disc closely matches that of expert ophthalmologists. AI-GS achieved a sensitivity of 0.9352 (95% CI 0.9277–0.9435) at 95% specificity. In real-world testing, sensitivity dropped to 0.5652 (95% CI 0.5218–0.6058) at ~0.9376 specificity (95% CI 0.9174–0.9562) for the standalone binary glaucoma classification model, whereas the full AI-GS network maintained higher sensitivity (0.8053, 95% CI 0.7704–0.8382) with good specificity (0.9112, 95% CI 0.8887–0.9356). The sub-models in AI-GS, with enhanced capabilities in detecting early glaucoma-related structural changes, drive these improvements. With low computational demands and tunable detection parameters, AI-GS promises widespread glaucoma screening, portable device integration, and improved understanding of disease progression.
17 July 2024 in Ophthalmology
29 March 2024 in Ophthalmology Glaucoma
To investigate sectoral differences in the relationship between intraocular pressure (IOP) dynamics during dark-room prone testing (DRPT) and visual field (VF) defect progression in primary open-angle glaucoma (POAG) patients. Retrospective, longitudinal study PARTICIPANTS: This retrospective study included 116 eyes of 84 POAG patients who underwent DRPT and had at least five reliable VF tests conducted over a more than two-year follow-up period. We excluded eyes with mean deviation worse than -20 dB or a history of intraocular surgery or laser treatment. Average total deviation (TD) was calculated in the superior, central, and inferior sectors of the Humphrey 24-2 or 30-2 program. During DRPT, IOP was measured in the sitting position, and after 60 minutes in the prone position in a dark room, IOP was measured again. The relationship between IOP change during DRPT, IOP after DRPT, and TD slope in each quadrant was analyzed with a linear mixed-effects model, adjusting for other potential confounding factors. TD slope in each quadrant, IOP change during DRPT, and IOP after DRPT RESULTS: IOP after DRPT and IOP change during DRPT were 18.16±3.42 mmHg and 4.92±3.12 mmHg, respectively. Superior TD slope was significantly associated with both IOP after DRPT (β=-0.28, p=0.003) and IOP change during DRPT (β=-0.21, p=0.029), while central (β=-0.05, p=0.595; β=-0.05; p=0.622) and inferior (β=0.05, p=0.611; β=0.01, p=0.938) TD slopes were not. DRPT might be a useful test to predict the risk of superior VF defect progression in eyes with POAG.