Leveraging AI-driven decision support systems to improve cataract surgery outcomes

A cataract is a progressive clouding of the eye's natural lens, which interferes with light passing through the eye and thereby results in blurred vision. Currently, there is no medical treatment to reverse or prevent this natural process. The cloudy lens must be physically removed and replaced with a clear synthetic implant called an intraocular lens (IOL). Surgical interventions to remove a cloudy lens in the eye include Cataract Surgery (CS) and Refractive Lens Exchange (RLE). IOL implantation, following either CS or RLE, is one of the most commonly performed eye surgical procedures in developed economies. Nowadays, the emphasis in CS and RLE has shifted from technical concerns about the safe removal of the cloudy lens to the minimisation of post-operative refractive errors, i.e., blurred vision, caused by IOL power miscalculations. Uncorrected refractive errors represent one of the world's leading causes of blindness and they are a significant contributor to the global burden of eye diseases. Refractive errors are known to be even greater and more unpredictable for eyes with uncommon characteristics. The affected population includes patients who have eyes with long or short axial length, patients with keratoconus condition, patients with prior laser refractive surgery as well as the growing senior patients’ population. Recently, new lens models, known as toric IOLs, have been designed with the aim to improve the refractive surgery outcomes for the aforementioned challenging eyes. However, this type of lenses required an accurate estimation of the two components (spherical and cylindrical) of its power, and this poses some challenges. Current industry standard formulae for calculating power of toric IOLs rely on classical formulae for IOL power calculation, which are designed to address only a specific (spherical) part of the refractive error. Hence, they fall short to accommodate the requirement of toric IOLs, and up to now there isn’t yet a satisfactory approach, capable of estimating both the adequate spherocylindrical power as well as the orientation of the corresponding astigmatic axis. To achieve the maximum potential of the visual benefit of toric IOLs, an accurate determination of their spherocylindrical power and the expected orientation of the astigmatic axis is of paramount importance. This requires innovative approaches, underpinned by new scientific understanding of the complex relationship between patient anatomical and physiological characteristics, IOL design as well as surgical techniques and instruments specific idiosyncrasies. This project aims to develop an innovative concept, which will address the deficiencies of current methods for calculating IOL power, leveraging the capabilities of Artificial Intelligence. The output of this project will culminate into a new methodology, which will be of significant benefit to the society through improved clinical quality of care and patients care experience; hence, a better quality of life for patients following CS or RLE. Ultimately, this will reduce the economic and social burden caused by eye diseases associated with uncorrected refractive errors while enhancing the UK's position as an international leader in eye-care research, service, and technologies.