Potential application of foveal structural measurements in treatment decision for retinopathy of prematurity: an OCT-based study

PURPOSE: To investigate foveal changes in infants with ROP not requiring treatment(nROP) and ROP infants needing treatment (tROP) using a handheld SD-OCT device. METHOD: We performed horizontal SD-OCT scans through the fovea in 156 eyes of 81 infants diagnosed with ROP. Foveal immaturity indices including the presence of inner retinal layers (IRL), absence of foveal outer nuclear layers widening (ONL) and attenuation of hyperreflective outer segment layers (OS), presence and type of cystoid macular edema (CME), epiretinal membrane (ERM), foveal pit depth (FPD), foveal pit width (FPW) and central foveal thickness (CFT) were calculated. The multivariate logistic regression model was used to predict the need for treatment based on OCT measurements. RESULTS: The shape of the foveolar pit was not significantly different among tROP and nROP groups (P-value = 0.287, Chi-square test). IRL extrusion was incomplete in both tROP and nROP groups (P-value = 0.0.41, Chi-square test). Nevertheless, the presence of thicker IRL was more frequent in the nROP group in comparison with the tROP group (100% vs.64.8%, P-value = 0.001). CME was observed in 29% of eyes in the tROP group and 40% of eyes in the nROP group; however, this difference was not statistically significant (P-value = 0.32, Chi-square test). ERM was detected in 15 (75%) and 84 (65.6%) eyes in the nROP and tROP groups, respectively (P-value = 0.39, Chi-square test). Multivariate logistic regression analyses showed that the need for treatment was significantly associated with gestational age (GA), CFT and FPD (P-values 0.001 and 0.002 respectively). CONCLUSIONS: This study demonstrated GA, foveal pit depth and the central foveal thickness could accurately predict the need for treatment with sensitivity, specificity, and diagnostic accuracy of 97%, 65% and 91.7% respectively.