Purpose Central serous chorioretinopathy (CSC) is a retinal disease that frequently shows resolution and recurrence with serous detachment of the neurosensory retina. Here, we present a deep learning analysis of subretinal fluid (SRF) lesion segmentation in fundus photographs to evaluate CSC. Methods We collected 194 fundus photographs of SRF lesions from the patients with CSC. Three graders manually annotated of the entire SRF area in the retinal images. The dataset was randomly separated into training (90%) and validation (10%) datasets. We used the U-Net segmentation model based on conditional generative adversarial networks (pix2pix) to detect the SRF lesions. The algorithms were trained and validated using Google Colaboratory. Researchers did not need prior knowledge of coding skills or computing resources to implement this code. Results The validation results showed that the Jaccard index and Dice coefficient scores were 0.619 and 0.763, respectively. In most cases, the segmentation results overlapped with most of the reference areas in the annotated images. However, cases with exceptional SRFs were not accurate in terms of prediction. Using Colaboratory, the proposed segmentation task ran easily in a web-based environment without setup or personal computing resources. Conclusions The results suggest that the deep learning model based on U-Net from the pix2pix algorithm is suitable for the automatic segmentation of SRF lesions to evaluate CSC. Translational Relevance Our code implementation has the potential to facilitate ophthalmology research; in particular, deep learning–based segmentation can assist in the development of pathological lesion detection solutions.

No abstract

Purpose: In the present study, a case of double fovea artifact on spectral-domain optical coherence tomography (SD-OCT) was reported. Case summary: A nine-year-old male presented with blurred vision of both eyes. His best corrected visual acuity (BCVA) was 20/20 in both eyes, and complete ophthalmologic evaluation including fundus examination and fundus photography revealed no abnormality in both eyes. He underwent SD-OCT imaging with the Cirrus HD-OCT. The Macular Cube 512 × 128 protocol of his right eye revealed an unusual pseudo-duplication of the fovea in the vertical meridian. The same protocol in his left eye also rendered a pseudo-duplication of two foveas in the vertical and horizontal meridians on the retinal thickness map. Re-examination with the same OCT system and protocol was performed two weeks later after the patient received counseling on fixation during the examination, and it revealed normal contour of the fovea in both eyes. Conclusions: Double fovea artifact seen in SD-OCT is a rare artifact that can possibly lead to misdiagnosis and inappropriate clinical treatment. Since the artifact was resolved with better fixation of the patient, repeating the scan with better patient compliance is necessary when such an artifact is encountered.

Purpose:To compare the clinical effectiveness of 1% Prednisolone acetate ophthalmic solution and 0.1% Bromfenac sodium hydrate ophthalmic solution on prevention of cystoid macular edema after cataract surgery. Methods: A retrospective chart review of 349 patients who received phacoemulsification with intraocular lens implantation in Severance Hospital from July 2013 to January 2016 was performed. In these patients, 192 eyes received 1% Prednisolone acetate ophthalmic solution, and 157 eyes were treated with topical 0.1% Bromfenac sodium hydrate ophthalmic solution. The incidence and severity of cystoid macular edema (CME) were evaluated by retinal foveal thickness on optical coherence tomography for patients who showed best corrected visual acuity (BCVA) less than 0.5 (log MAR ≥ 0.3). Results: There was no significant difference between the two groups in age (p = 0.708), sex (p = 0.977), or the side of operated eye (p = 0.443). The two groups showed BCVA 0.04 ± 0.09 (Steroid group) and 0.03 ± 0.07 (nonsteroidal anti-inflammatory drug [NSAID] group) at 1 month after the surgery and the difference was not significant (p = 0.947). One eye in the topical steroid group had cystoid macular edema, and 3 eyes in the steroid group showed elevated intraocular pressure (IOP) over 30 mm Hg. There were no IOP elevations or macular edema in the NSAID group. Conclusions:The results showed that 0.1% Bromfenac sodium hydrate ophthalmic solution had a similar effect to 1% Prednisolone acetate ophthalmic solution on preventing CME after cataract surgery. This indicates that topical NSAID can be considered along with topical steroids in order to prevent CME after cataract surgery.

Purpose:To compare the outer retinal thickness in normal fellow eyes of patients with unilateral age-related macular degeneration (AMD) and normal control eyes. Methods: We retrospectively reviewed the medical records of 60 patients with unilateral exudative AMD including polypoidal choroidal vasculopathy and 60 normal controls. Spectralis optical coherence tomography was performed in the normal fellow eyes of patients with unilateral AMD and in the normal group. The thicknesses between the retinal pigment epithelium (RPE) line and the cone outer segment tips (COST) line, between the COST line and the photoreceptor inner segment/outer segment (IS/OS) line, and between the IS/OS line and the external limiting membrane (ELM) line were measured at the fovea in both groups. Results:The thickness between the RPE and COST lines was 32.4 ± 3.0 μm in normal fellow eyes of patients with unilateral AMD and 35.3 ± 3.5 μm in the normal group (p < 0.001). Total retinal thickness, thicknesses between the COST and the IS/OS lines and the IS/OS and the ELM lines in fellow eyes were not significantly different from those of normal eyes (p = 0.126, 0.615, 0.874). There was no significant difference in total retinal thickness or each outer retinal thickness measured in normal fellow eyes between patients with neovascular AMD and polypoidal choroidal vasculopathy. Conclusions:The thickness between the RPE and the COST lines was thinner in the fellow eyes of patients with unilateral AMD than in the normal eyes. We suggest that less thickness between the RPE and COST lines might indicate a greater risk of AMD.