Prediction of Postoperative Visual Outcome in Patients with Idiopathic Epiretinal Membrane

Abstract

<b><i>Purpose:</i></b> The aim of this study was to estimate the association between preoperative characteristics in subjects with idiopathic epiretinal membrane (ERM) and visual acuity improvement after vitrectomy and create an algorithm for predicting postoperative visual outcome. <b><i>Methods:</i></b> In a retrospective, cross-sectional study, we included adults with idiopathic ERM and excluded subjects with low-quality scans, other ocular conditions, and previous surgery except cataract surgery. Baseline characteristics were extracted from medical files, spectral-domain OCT, and OCT angiography. Visual improvement was expressed as a binary variable. <b><i>Results:</i></b> Fifty-four subjects were included in the study. Three months postoperatively, 30 subjects improved, 10 remained stable, and 14 deteriorated. Spearman correlation showed no correlation between variables and visual acuity improvement (<0.39). Reduced dimensionality showed that baseline visual acuity, lens status, foveal aspect, spherical equivalent, and 2 interactive variables including foveal aspect and lens status have the strongest effect on improvement. Five-fold logistic regression based on these variables provided a model with AUC 0.9 ± 0.06. <b><i>Conclusion:</i></b> No variable has a direct predictive role on visual acuity improvement; however, baseline visual acuity, lens status, foveal aspect and spherical equivalent, when combined, provide a predictive model that could serve as a tool for more informed decisions.