Purchase this article with an account.
William R. Freeman, Jae Suk Kim, Jay Chhablani, Kathrin Hartmann, Lingyun C. Cheng; Predictive Value of SD OCT Morphology on Surgical Findings in Macular Pucker. Invest. Ophthalmol. Vis. Sci. 2011;52(14):2206.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the use of spectral domain OCT to predict surgical findings and ease of membrane peeling in eyes undergoing pars plana vitrectomy for epiretinal membrane and related retinal surface disease.
A total of 25 consecutive eyes undergoing surgery for ERM were studied. All had undergone videography of ERM surgery as well as careful raster scanning of the posterior pole with SD OCT as well as fluorescein angiography (Heidelberg Spectralis). Videos of the ERM peeling were studied by an independent observer. Separately, the surgeon’s intraoperative impressions of ease of membrane removal were noted as they were recorded on a surgical sketch form. Masked reviewers graded epiretinal membrane adhesion by analysis of multiple SD OCT scans of the macula ERM was graded for peeling ease on a scale of 1-3 in the four quadrants of the macula and in the fovea. The presence of fibrillary changes of the retinal nerve fiber layer were also graded by quadrant. Similarly, a masked observer and the surgeon graded ease of peeling in each of the five macular locations on an ordinal scale.
Concordance of graders of OCT was extremely high (Kappa=0.93 and Kappa=1). Generalized estimating equation modeling showed a strong correlation of ease of surgical removal with ERM adhesion grading on OCT (p=0.0286) as well as presence of fibrillary change on OCT (p=0.0043). Both ERM grading and fibrillary change had additive predictive value to ease of ERM removal. With fibrillary effect in the model, ERM with OCT grading of 1 had a higher probability to be in the easiest peeling category compared with ERM which had a OCT grading 2 (8 fold higher, p=0.01) or grade 3 (27 fold higher, p<0.001). Similarly, with ERM OCT grading in the model, ERM with no fibrillary changes had 116 times probability to be in the easiest peeling category compared to the ERM having fibrillary changes on OCT (p=0.0017).
Examination of OCT scans for gaps between ERM and ILM helps predict ease of membrane peeling and the best location to start the peel. OCT visible fibrillary change of the RNFL is associated with more difficulty in ERM peel and likely means a strong adhesion of the ERM to the retinal surface was the cause of the fibrillary change.
This PDF is available to Subscribers Only