Abstract
Introduction Several optical coherence tomography (OCT) biomarkers for postsurgical success in
patients with idiopathic epiretinal membranes (iERMs) have been described in single
predictor analyses. The aim of this study was to assess the reliability of diagnosis
of OCT biomarkers and to calculate their impact on postsurgical visual acuity by applying
multiple
regression analysis.
Methods This retrospective study included patients scheduled for pars plana vitrectomy with
membrane peeling for iERMs. Presurgical OCT scans were analyzed for the presence of
OCT
biomarkers by two reviewers. Intra- and interobserver reliability was calculated with
the Kuder-Richardson 20 test.
Results Among the 136 patients included, intra-/interobserver reliability of the OCT biomarkers
was 0.89/0.70 for ectopic inner foveal layer (EIFL), 0.82/0.53 for disorganization
of
retinal inner layer (DRIL), 0.93/0.89 for intraretinal cystoid changes, 0.84/0.78
for alterations of the ellipsoid zone, 0.84/0.72 for cotton ball sign, 0.68/0.58 for
hyperreflective foci,
0.55/0.47 for epimacular membrane rips (EMM-rips), and 0.87/0.60 for retinal contraction.
Only DRIL and central subfield macular thickness (CMT) were significant predictors
for BCVA 3 months
after surgery in a multiple regression analysis.
Conclusion Intraobserver reliability of OCT biomarkers was higher than interobserver reliability,
reflecting both clinical and grading experience with respect to OCT biomarker
evaluation. DRIL and CMT were shown to be significant predictors in a multiple regression
analysis, with potential for negative influence on final BCVA after vitrectomy with
membrane peeling
for iERMs.
Key words vitrectomy with membrane peeling - OCT biomarker - epiretinal membrane
