Macular Imaging with Spectral-Domain Optical Coherence Tomography

The scan protocol consisted of a high-density acquisition of 25 consecutive horizontal B-scans centered on the fovea. B-scans were equally spaced 60 μm apart, covering a central foveal field of 20 × 5 degrees. Lesions in SD-OCT imaging were evaluated by 2 independent observers (M.P.D., C.C.G.). Macular thickness was obtained by automatic segmentation between the inner retinal border and the outermost reflective band (retinal pigment epithelium/Bruch membrane complex). The thickness of the ONL and of the PROS layer required manual segmentation on consecutive B-scans. The PROS thickness was the distance between the retinal pigment epithelium/Bruch membrane complex and the photoreceptor inner-outer segment junction. The thickness of the ONL was measured between the high reflective line of the ELM and the outermost reflection of the outer plexiform layer. The type and number of lesions detected in SD-OCT were as follows: hyperreflectivity spots in the outer nuclear layer, disruption of the external limiting membrane, disruption of the IS/OS junction, and an irregular hyporeflectivity in the PROS.

Fundus Autofluorescence

FAF photography was performed using a band-pass filter for the excitation light centered at 550 nm (bandwidth, 535−585 nm) and a matched barrier filter centered at 665 nm (bandwidth, 615−715 nm). The FAF topographic profile generally follows the distribution of lipofuscin in retinal pigment epithelium (RPE) cells, and a decrease in autofluorescence generally indicates a corresponding loss of or damage to the involved RPE cells. Fundus autofluorescence abnormalities seen in this series were recorded at 4 degrees and 12 degrees.

Microperimetry

The recently developed Spectral OCT/SLO (Opko/Ophthalmic Technologies Inc) provides a perimetry of the central visual field with a scale of 0 to 20 decibels (dB) and can detect microscotomata; it also superimposes the microperimetry map onto the retinal thickness map. An eye tracking system is incorporated to compensate for eye movements. In this study, we used Goldmann III stimuli with a presentation time of 200 ms and a 4−2−1 staircase strategy. Twenty-eight locations were tested on a circular grid covering the 12 degrees of the central area. The 4-degree central sensitivity was also recorded. The examination was performed before pupil dilation and started on the fellow eye to limit the learning effect as all patients were inexperienced in automated perimetry. Fixation stability and response reliability (false-positive and false-negative rates) were monitored during the examination. Fixation was “stable” if more than 75% of the fixation points were gathered within a 2-degree central area. Fixation was “unstable” if less than 75% of the points were located inside a 4-degree central circle. Otherwise, fixation was regarded as “relatively unstable.” All our patients had less than 30% false-positive and false-negative responses.

For statistical analysis, preoperative and postoperative BCVA were converted to logMAR value. Patients were split into 2 groups according to their postoperative BCVA: 18 patients in group A with good BCVA (less than 0.3 logMAR) and 12 patients in group B with poor BCVA (0.3 logMAR or higher). Comparisons were performed with the Fisher exact test for dichotomous data. A nonparametric Mann-Whitney test was used for continuous variables such as age or visual acuity. The Spearman rank correlation coefficient was used as a measure of association between non-normally distributed variables, such as retinal sensitivity results, retinal thickness, or the number of lesions in the photoreceptor layer. All analyses were conducted using Prism 5.01 (Graphpad Software, La Jolla, California, USA). The level of statistical significance was set at P < .05 and the tests were 2-tailed.