Case Selection

A retrospective chart review was performed of consecutive patients with resolved CSC at Korea University Ansan Hospital between March 1, 2009 and September 30, 2012. Each patient had a documented episode of CSC diagnosed by fluorescein angiography. CSC resolution was confirmed with SD OCT. Eyes were included if at least 4 months had elapsed since confirmation of CSC resolution by SD OCT and either short-wavelength AF or near-infrared AF image findings were abnormal. If the exact disease onset was unknown, symptom onset was used instead. When AF images were examined, SD OCT and fluorescein angiography were performed again to verify complete attachment of the retina and no fluorescein leakage. Eyes with other intraocular diseases (ie, age-related macular degeneration, polypoidal choroidal vasculopathy, epiretinal membrane, and diabetic retinopathy) or those with a history of ocular surgery were excluded. Eyes with a history of cataract surgery were not excluded.

Fundus Autofluorescence and Fluorescein Angiography

AF and fluorescein angiography images were taken using a confocal scanning laser ophthalmoscope (SLO) (Heidelberg Retina Angiograph 2 [HRA]; Heidelberg Engineering, Heidelberg, Germany) with a 30-degree field of view. The image resolution was 768 × 768 pixels. Short-wavelength AF and near-infrared AF images were obtained simultaneously. For short-wavelength AF imaging, a wavelength of 488 nm was used for excitation, and emitted light was detected above 500 nm with a barrier filter. Near-infrared AF images of ocular fundi were obtained at an excitation wavelength of 787 nm using a barrier filter for the detection of emitted light above 810 nm. For both short-wavelength AF and near-infrared AF, up to 100 single images were averaged (4.7 frames/s) depending on patient fixation using the HRA mean algebraic reconstruction technique algorithm to obtain a high-quality mean image. Fluorescein angiography images were selected around 40 seconds after dye injection.

Abnormal AF lesions were categorized as either dot-like or confluent according to their configuration. Confluent autofluorescence was defined as a geographic area of increased or decreased autofluorescence, and dot-like autofluorescence as a granular, coarse lesion. The AF composite inside the lesion was classified as hyper-, hypo-, or mixed-AF type ( Figures 1-4 ). Mixed-type was defined as an abnormality with the coexistence of hyper- and hypo-AF. Lesions were qualitatively classified as hyper-, hypo-, or mixed-AF type lesions.

A 39-year-old male patient (Group 1) complained of decreased vision in his right eye in the previous 3 years. These images were taken 1 year after the resolution of central serous retinopathy. Best-corrected visual acuity (logMAR) improved from 0.52 to 0.00. (Top left) The short-wavelength fundus autofluorescence image looked normal. (Top middle) The near-infrared fundus autofluorescence image showed dot-like type hypo-autofluorescence; (Top right) on the fluorescein angiography image, we observed a hyperfluorescent area corresponding to the hypo- near-infrared autofluorescence lesion. (Bottom) On the spectral-domain optical coherence tomography raster scan, the ellipsoid portion of the inner segments (EPIS) looked normal. 1, external limiting membrane; 2, EPIS (previously referred to as the junction between the photoreceptor inner and outer segments); 3, contact cylinder (junction between photoreceptor tips and apical processes of retinal pigment epithelium); 4, retinal pigment epithelium.

A 42-year-old male patient (Group 2) complained of decreased vision in his right eye in the previous year. These images were taken 6 months after the resolution of central serous retinopathy. Best-corrected visual acuity (logMAR) improved from 0.40 to 0.00. (Top left) Short-wavelength fundus autofluorescence and (Top middle) near-infrared fundus autofluorescence images revealed a geographic heterogeneous mixed-type autofluorescence lesion. (Top right) The fluorescein angiography image showed a hyperfluorescent lesion corresponding to the hypo-autofluorescence lesion. (Bottom) On the spectral-domain optical coherence tomography raster scan image, the retinal pigment epithelial layer appeared to have collapsed, whereas the ellipsoid portion of the inner segments maintained its contour.

A 43-year-old male patient (Group 3) complained of decreased vision in his right eye in the previous 3 years. These images were taken 1 year after the resolution of central serous retinopathy. Best-corrected visual acuity (logMAR) improved from 0.70 to 0.30. Images show a lesion with confluent geographic-type hypo–fundus autofluorescence on (Top left) short-wavelength fundus autofluorescence and (Top middle) near-infrared fundus autofluorescence. (Top right) The fluorescein angiography image also showed a hyperfluorescent lesion corresponding to the hypo-autofluorescence lesion. (Bottom) On the spectral-domain optical coherence tomography raster scan image, the outer nuclear and ellipsoid portion of the inner segments layers appeared to have collapsed.

A 41-year-old male patient (Group 3) complained of decreased vision in his left eye in the previous 6 years. These images were taken 4 years after the resolution of central serous retinopathy. Best-corrected visual acuity (logMAR) improved from 0.52 to 0.15. (Top left) Short-wavelength fundus autofluorescence and (Top middle) near-infrared fundus autofluorescence images revealed confluent scar-like hyper- and hypo-autofluorescence lesions; we categorized this as mixed-type autofluorescence. (Top right) Fluorescein angiography revealed a hyperfluorescent window defect; (Bottom) on the spectral-domain optical coherence tomography image, the ellipsoid portion of the inner segments and retinal pigment epithelial layer appeared to have collapsed.

Optical Coherence Tomography

SD OCT was performed using the Spectral OCT/SLO system (OPKO/OTI, Miami, Florida, USA). All SD OCT images included line scans (horizontal and vertical B-scans), raster scans, and topographic mapping. Line scans were performed horizontally and vertically and the scan image was created by averaging up to 20 B-scans (1024 A-scans per B-scan) of 9 mm in length. The raster scan was performed using 32 B-scans (1024 A-scans per B-scan) of a 9 × 9-mm area. Topographic mapping consisted of 256 horizontal A-scans × 256 vertical lines across a 5 × 5-mm area. The line scans, raster scans, and topographic mapping were performed on each eye.

To compare the abnormal AF lesion with the OCT scan, the OCT raster scan image located on the middle area of the AF lesion was selected. If there was no proper OCT scan line corresponding to the middle location because of the width limitation of the raster scan, we selected the most abnormal-looking cut image within the lesion. In the OCT image, we focused on abnormalities in the outer retina. Reflective line integrities of the ellipsoid portion of the inner segments (EPIS, previously known as the junction between the inner and outer segments of the photoreceptors) on the OCT scan were assessed qualitatively. We assigned eyes with a defective EPIS signal to the EPIS damaged group.

To compare visual acuity according to EPIS and AF abnormalities, we also took the fovea center thickness into consideration. To measure fovea center thickness, central retinal thickness was assessed at the macular center (1 mm diameter) using topographic mapping. The central retinal thickness was generated automatically during OCT analysis.

Image Analysis

Patients were grouped according to the results of near-infrared AF, short-wavelength AF, SD OCT, and fluorescein angiography. In each group, we compared the abnormality patterns on short-wavelength AF and near-infrared AF with the retinal changes on SD OCT and fluorescein angiography.

Statistics

An independent t test or a Mann-Whitney test was performed to compare visual acuity and central retinal thickness between subgroups. Comparison of AF characteristics (configuration and AF composite) among subgroups and the relationship between AF and OCT changes were assessed by means of χ ² tests. All statistical procedures were performed using SPSS version 17.0 (SPSS Inc, Chicago, Illinois, USA). A P value < .05 was considered statistically significant.