Purchase this article with an account.
Marwan Abdulaal, Michael M Lin, Migil Ledesma, Cloyd Pitoc, Paolo S Silva, Michael Cheney, Jan Lammer, Tahera Doctor, Lloyd Paul Aiello, Jennifer K Sun; Reproducibility of Disorganization of the Retinal Inner Layers (DRIL) Grading Across Spectral Domain Optical Coherence Tomography (SDOCT) Model and Scan Types in Eyes with Diabetic Macular Edema (DME). Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):5959.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate reproducibility of DRIL evaluation in eyes with DME imaged by 2 SDOCT instruments (Zeiss Cirrus HD-OCT 4000D and Heidelberg Spectralis 6 mode) with high speed and high resolution settings.
A cross sectional study enrolled eyes with current or recently resolved (<3 month) center-involved DME (ciDME) defined using gender/machine specific values equivalent to time domain CST>250µm. For each eye, 2 Cirrus (Macular Cube 512x128 [Cube] and HD 5 Line Raster [5 Line]) and 2 Spectralis (20x20°, 49 B-scans, 16 ART Mean: high res [HR] and high speed [HS] settings) scans were acquired. The 1 mm area centered on the fovea was graded using customized MATLAB software for 7 B-scan substacks in HR, HS and Cube scans and all 5 B-scans for 5 Line scans. Presence and extent of DRIL, intraretinal cysts and cone outer segment tip (COST) visibility were determined. DRIL causes were assessed as either generalized inner retinal layer boundary blurring (blur), presence of intraretinal cysts crossing boundary lines (cysts) or hyperreflective foci (HR) obliterating layer demarcations.
In 24 eyes of 18 patients, mean±SD CST was 360±94μm, 67% (N=16) had current and 33% (8) had recently resolved ciDME. Mean age was 61±15 years and 47% were female. Intraclass correlation coefficients (ICC) and 95% confidence limits (CL) between HR scans and each of the other scan types for extent of DRIL, DRIL cause, intraretinal cyst boundary area (horizontal x vertical extent), and COST visibility are reported in Table 1. ICCs were highest for overall DRIL extent, DRIL caused by blur or cysts, and cyst boundary area. Grading was most reproducible between HR and HS scans, followed by HR and Cirrus Cube scans.
Reproducibility for SDOCT parameters of DRIL and intraretinal cysts was high overall and excellent across OCT models and scan types, especially between Spectralis HR and HS scans. These data suggest that evaluation of DRIL and other SDOCT variables is feasible using these different OCT models and scan types in eyes with DME.
This PDF is available to Subscribers Only