Abstract
Study design-Retrospective,Purpose:To identify pre-operative& risk factors affecting outcomes of open globe injuries using AI Methods:A review of 1324 eyes who underwent surgical repair for open globe injury over past 10 years at a tertiary referral eye center was done.Based on vision recovery post surgery,good/bad outcomes were categorised.Pre-operative visual acuity(VA),extent of wound,RAPD,cataract,vitreous loss,age,time since injury,distance to hospital were analysed.AI model was built to assess these factors Results:Median age was 38.2 years with male predilection(72%).Pediatric age constituted 39%. Age,time to hospital,extent of wound involving visual axis,vitreous loss,cataract had highest information gain.AI model accurately predicted outcomes in 92% & AUC of 0.912 Conclusion:Identifying risk factors affecting outcomes of open globe injuries can aid in better stratification & management.AI model helps in addressing loopholes in referrals & treatment,thereby optimizing outcomes.
