AI-Powered Prediction System for Pupil Morphology in Critical Stages of Cataract Surgery

Cataract surgery is the primary treatment for cataracts, a leading cause of preventable blindness worldwide. Maintaining sufficient and stable pupil dilation is crucial for a successful procedure, as pupillary instability increases the risk of surgical complications. Monitoring pupil morphology changes during surgery is essential for optimizing surgical decision-making. Predicting pupil morphology changes intraoperatively, especially in the later surgical phases, can help surgeons anticipate potential challenges, adjust techniques, and implement appropriate measures. However, both monitoring and predicting pupil morphology in cataract surgery remain challenging due to high variability among procedures. In this study, we introduce a novel AI-powered prediction system designed to accurately monitor and predict pupil morphology changes in critical later stages of surgery using data collected in the early stages. To develop and validate the system, we utilized 500 cataract surgery video recordings from human patients. Experimental results demonstrate that the proposed system effectively predicts pupil morphology, highlighting its potential as a tool for providing intraoperative guidance to cataract surgeons.