Abstract
Purpose::
To describe the three-dimensional imprint of choroidal neovascular membranes on the retinal pigment epithelium using segmentation algorithms of eyes studied with spectral domain optical coherence tomography (SD-OCT).
Methods::
All SD-OCT scans from patients at the Bascom Palmer Eye Institute diagnosed with exudative age-related macular degeneration and polypoidal choroidal vasculopathy were reviewed. Cases with a clearly defined choroidal neovascular membrane on the acquired B-scans were analyzed using a macular segmentation algorithm. The resulting model of the retinal pigment epithelium (RPE) could then be compared to registered fundus photographs, fluorescein angiograms, and when applicable, indocyanine green angiograms.
Results::
Rapid acquisition of sequential adjacent B-scans through the macula allows reconstruction of a high-density volume of tissue that can be analyzed in multiple planes. A segmentation algorithm can then be used to accurately construct a continuous model of the RPE.Distinct morphological subsets of choroidal neovascular membranes were identified. In several eyes with type I choroidal neovascular membranes a three-dimensional representation of the upward displacement of the RPE was obtained. The contour of the segmentation has a steep sharp edge at its boundaries, and is tubular and undulating within. This is consistent with vascular elements, and is easily differentiated morphologically from a pigment epithelial detachment, which is seen as a gradually sloping and uniform structure. The appearance of neovasculature in the segmentation algorithm was corroborated by registered fundus images and angiography. In a case of polypoidal choroidal vasculopathy, the contour of the plaque that is present on indocyanine green angiography is accurately reproduced as an edge of anteriorly-displaced RPE in the SD-OCT segmentation algorithm.
Conclusions::
Segmentation algorithms of SD-OCT data of eyes with exudative age-related macular degeneration can yield accurate representations of underlying choroidal neovascular membranes. These models may help further categorize choroidal neovascular membranes into distinct clinical subtypes and may aid in their quantitation.
Keywords: choroid: neovascularization • retina • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)