Artificial intelligence-driven imaging in sickle cell retinopathy: Advancing early detection, risk stratification, and precision ophthalmology

Sickle cell disease (SCD) is a disorder of genetically inherited red blood cells (RBCs), wherein it is manifested clinically as sickling of RBCs, due to polymerization of hemoglobin S. SCD patients frequently face RBCs clumping, blocking of blood vessels and capillaries along with persistent hemolysis.Sickle cell retinopathy (SCR), is a SCD associated disorder in the peripheral retina, responsible for several features like vasoocclusion, ischemia, vitreous hemorrhage and neovascularization within the retina and adjoining regions.While these are manageable features, some severe cases report retinal detachment and ultimately vision impairment in patients.As early signs of SCR are not usually manifested visually, its early diagnosis becomes the limiting factor for devising effective management strategies in patients.Artificial intelligence (AI) assisted deep learning models have very recently being utilized and provides a promising alternative in assisting early diagnosis, staging, and monitoring disease progression.Traditional retinal imaging methods like fundus photography, fluorescein angiography, optical coherence tomography and optical coherence tomography angiography scans have always assisted ophthalmologists to make clinical decisions for patients.Recently, these methods were being used to train AI models for easy and efficient analysis of such scans.As a growing field, works in this area have transitioned our understanding of image analyses and promises a fast, objective and efficient interpretation.This review discusses; how traditional methods captures structural damages to peripheral retina and summarizes recent research work done in this field to integrate AI driven models based on deep learning and machine learning.We summarize here AI based strategies for SCR screening and diagnosis, while underlining their clinical applicability, utilities and future translational promises towards ascertaining personalized ophthalmology.