Patient demographics, clinical and imaging data were retrospectively collected on a consecutive group of AFVD cases. All patients were treated and followed in the retina service of the Department of Ophthalmology at the Hadassah-Hebrew University Medical Center in Jerusalem, Israel, between January 2010 and January 2017. Patients signed an informed consent form and the study was approved by the institutional ethics committee.
The EZ was segmented and evaluated using OCT (Spectralis - Heidelberg Engineering, Germany) scans from 92 eyes (n = 57 patients) with AFVD, which were included in a previous report on the long term follow up of the disease [11]. Inclusion and exclusion criteria for this cohort were previously described [11]. Briefly, patients with AFVD in at least one eye at any stage were included in the study. Vitelliform lesions associated with any other pathology, such as Best’s disease, vitreomacular traction (VMT) or epiretinal membrane were excluded. Scans with poor image quality which hinder proper segmentation (shadow artifact, motion artifact, and noise) were excluded.
Four consecutive OCT scans from each study eye with an interval of 3 to 6 months between the visits were analyzed. For automated multi-layer retinal segmentation we used the 6.3a software version of the Spectralis SD-OCT. Based on the reflectivity of the different retinal layers, the software produced 13 lines representing the boundaries of the retinal layers: internal limiting membrane (ILM), retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), external limiting membrane (ELM), first photoreceptor layer (PR1), second photoreceptor layer (PR2), retinal pigment epithelium (RPE), bruch's membrane (BM), and choroid (CHO). However, the software does not measure the EZ. Based on the segmentation generated by the software we identified the EZ and measured its thickness (Fig. 1). To that end, we manually adjusted the lines that represent the RPE to the inner edge of the hypo-reflective area (zone 12) [14], and the outer retinal layer line to the hypo-reflective line of the myoid zone of the photoreceptors (Fig. 1). Finally, we subtracted the RPE layer thickness from the thickness of the outer retinal layer per the automatic measurement of the Spectralis software and got the thickness of the isolated ellipsoid zone layer. We measured the thickness in two areas of the macula: at the center of fovea (CF), and at the edge of the foveal avascular zone (FAZ) 500 microns nasal to the fovea (FE) (Fig. 1). The nasal edge of retina is the thickest part so its measurement minimalizes measurement errors [36–37].
In addition to segmentation of the retina layers, the following features were also analyzed for each visit: (1) stage of the vitelliform lesion (vitelliform, vitelliruptive, pseudohypopyon and atrophy); (2) integrity of the EZ which was graded as intact, disrupted or absent; (3) presence of sub or intra retinal fluid (SRF, IRF), retinal pigment epithelial detachment (PED) or drusen (4) vitreoretinal interface abnormalities including ERM, VMT and posterior vitreous detachment (PVD).
The entire procedure was performed by two experiences graders (S.Y. and L.T.) in order to investigate the repeatability of all measurements. All values were then averaged to perform the statistical analysis.
Statistical analyses were performed using SPSS (version 25.0; IBM Corp., Armonk, NY). In order to evaluate a single OCT scan as a separate case, we used a repeated measures ANOVA model to evaluate the association of the change over time in the visual acuity with the progression of the stages. Comparing stage between different time points was performed using the McNemar-Bowker's test. To compare quantitative variables (e.g. months of follow-up) between two independent groups (change in stage, etc.) the Mann-Whitney non-parametric test was applied. The ANOVA test with post hoc comparisons with the Dunnett correction was used for comparing quantitative variables between three independent groups (e.g. visual acuity by thickness in fovea or in distance from fovea, in three categories). The Chi-square and the Fisher's exact tests were used for evaluating the association between two categorical variables (stage, etc.), and the linear by linear association test was used for assessing trend. Finally, Intraclass correlation coefficients (ICC) between graders was calculated for the EZ thickness and integrity. All tests applied were two-tailed and a p-value of 5% or less was considered statistically significant.