Abstract
Background/Objectives: Ptosis surgery corrects drooping upper eyelids, improving function and esthetics. Traditional methods like marginal reflex distance (MRD) and palpebral fissure height (PFH) offer limited one-dimensional measurements. This study evaluates AI-based corneal exposure ratio (CER) analysis, a two-dimensional approach, compared to manual ImageJ methods for assessing ptosis surgery outcomes. Methods: In this prospective study, 100 eyes from 50 patients were analyzed using both methods. AI-based CER measurements were compared to manual ImageJ measurements for reliability and accuracy. Results: AI-based CER measurements were comparable to manual ImageJ, with high reliability (ICC 0.992, 0.985). Preoperative CER was 55.34% (manual) and 55.79% (AI), increasing to 75.92% (manual) and 75.84% (AI) postoperatively. The AI tool showed minimal bias and high repeatability (ICC 1.000), offering faster automated measurements. Conclusions: AI-based CER analysis matched manual methods in accuracy but provided significant efficiency advantages, making it suitable for clinical use. Limitations include a small homogeneous sample size and reliance on 2D imaging, which may not fully capture three-dimensional changes. Further studies are recommended to enhance generalizability and precision.