Abstract
Purpose: :
Conventional optical coherence tomography (OCT) analysis algorithms have been shown to frequently misidentify retinal boundaries in patients with choroidal neovascularization. To determine the extent, significance, and cause of these errors, corrected retinal thickness maps generated by reading center software (termed OCTOR) were compared with the Stratus OCT output.
Methods: :
Raw exported Stratus OCT image data from 55 consecutive cases received at the Doheny Image Reading Center was used in this analysis. The OCTOR reading center software was used to identify the inner and outer retinal boundaries in each of six radial line scans to generate retinal thickness maps and thickness measurements for each of the nine standard subfields. Retinal thickness measurements in the four middle and four outer subfields were summed and averaged to generate mean thicknesses for the "middle" and "outer rings." Retinal thickness measurement for the foveal center, foveal subfield, middle ring, and outer ring were compared between the OCTOR analysis and the Stratus output. Mean, median, and % differences were calculated.
Results: :
The mean % difference in retinal thickness between OCTOR and Stratus was 18% for the foveal center, 10% for the foveal subfield, 8% for the middle ring, and 7% for the outer ring. Nearly 1/3 (18/55) of the cases showed a difference of 20% or more in the Stratus analysis versus OCTOR–corrected measurement of the foveal center thickness, with a maximum difference of 54%. Review of cases with large discrepancies between Stratus and OCTOR measurements revealed that Stratus errors were frequently due to the presence of tissue or fluid in the subretinal space and variable inclusion of this material within the outer retinal boundary by the Stratus algorithm. OCTOR analysis allowed separate quantification of the retinal thickness and the subretinal space.
Conclusions: :
Manual correction of the retinal boundaries by the OCTOR software appears to be of value in patients with choroidal neovascular disease. OCTOR analysis permits analysis of all subfields, and also allows quantification of the subretinal space.
Keywords: imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • choroid: neovascularization