Pseudoaveraging for denoising of OCT angiography: a deep learning approach for image quality enhancement in healthy and diabetic eyes.

BACKGROUND: This study aimed to develop a deep learning (DL) algorithm that enhances the quality of a single-frame enface OCTA scan to make it comparable to 4-frame averaged scan without the need for the repeated acquisitions required for averaging. METHODS: Each of the healthy eyes and eyes from diabetic subjects that were prospectively enrolled in this cross-sectional study underwent four repeated 6 × 6 mm macular scans (PLEX Elite 9000 SS-OCT), and the repeated scans of each eye were co-registered to produce 4-frame averages. This prospective dataset of original (single-frame) enface scans and their corresponding averaged scans was divided into a training dataset and a validation dataset. In the training dataset, a DL algorithm (named pseudoaveraging) was trained using original scans as input and 4-frame averages as target. In the validation dataset, the pseudoaveraging algorithm was applied to single-frame scans to produce pseudoaveraged scans, and the single-frame and its corresponding averaged and pseudoaveraged scans were all qualitatively compared. In a separate retrospectively collected dataset of single-frame scans from eyes of diabetic subjects, the DL algorithm was applied, and the produced pseudoaveraged scan was qualitatively compared against its corresponding original. RESULTS: This study included 39 eyes that comprised the prospective dataset (split into 5 eyes for training and 34 eyes for validating the DL algorithm), and 105 eyes that comprised the retrospective test dataset. Of the total 144 study eyes, 58% had any level of diabetic retinopathy (with and without diabetic macular edema), and the rest were from healthy eyes or eyes of diabetic subjects but without diabetic retinopathy and without macular edema. Grading results in the validation dataset showed that the pseudoaveraged enface scan ranked best in overall scan quality, background noise reduction, and visibility of microaneurysms (p < 0.05). Averaged scan ranked best for motion artifact reduction (p < 0.05). Grading results in the test dataset showed that pseudoaveraging resulted in enhanced small vessels, reduction of background noise, and motion artifact in 100%, 82%, and 98% of scans, respectively. Rates of false-positive/-negative perfusion were zero. CONCLUSION: Pseudoaveraging is a feasible DL approach to more efficiently improve enface OCTA scan quality without introducing notable image artifacts.

Preretinal hemorrhage preventing subretinal involvement in multilayer macular hemorrhages: coincidence or plausible relationship?

BACKGROUND: To report four cases with interesting anatomical presentations of multilayer macular hemorrhages with preretinal hemorrhage possibly preventing subretinal involvement of the macular area. CASES PRESENTATION: Observational study of four patients presenting with macular hemorrhages. RESULTS: Four patients with multilayer macular hemorrhage due to different causes, presented with a halo-shaped submacular hemorrhage coincident with the preretinal hemorrhage borders. After resolution, in all cases, the macular area underneath the preretinal hemorrhage was found to be spared. CONCLUSION: We hypothesized that an extensive preretinal hemorrhage can exert a mechanical force pushing the subretinal hemorrhage towards the periphery, consequently protecting the macular area.

SWEPT-SOURCE OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY IN SMALL CHOROIDAL MELANOMAS AND CHOROIDAL NEVI.

PURPOSE: To evaluate the use of swept-source optical coherence tomography angiography to detect distinct vascular features in small choroidal melanomas and choroidal nevi. METHODS: Patients with a choroidal nevus or a treatment-naïve choroidal melanoma were imaged with color fundus photography, ultrasound, and swept-source optical coherence tomography angiography (12 × 12 mm). High-risk features including overlying fluid, orange pigment, shaggy photoreceptors, acoustic hollowness, depth >2 mm, and basal diameter >5 mm were assessed. Optical coherence tomography angiography vascular markers included: choroidal vessel visualization, choroidal vessel depth, and choriocapillaris flow signal, assessed qualitatively by comparison with surrounding, unaffected choriocapillaris. RESULTS: Twenty-nine lesions were included in this study, seven flat choroidal nevi, 17 elevated choroidal nevi, and 5 choroidal melanomas. Distinct vascular patterns were noted between flat nevi, elevated nevi, and small choroidal melanomas. Choroidal melanomas displayed two types of vasculature: "nevus-like" vasculature with straight parallel vessels and complex vasculature with vascular loops and crosslinking. Visualized choroidal vessels were significantly deeper in melanomas (110 µm) than elevated (84 µm) or flat nevi (70 µm). In a size-matched subanalysis of 5 elevated choroidal nevi and 5 choroidal melanomas, choroidal melanomas had increased mean choroidal vessel depth (P = 0.015), deepest choroidal vessel visualized (P = 0.034), and presence of a deep choroidal vessel >155 µm (P = 0.048). CONCLUSION: Swept-source optical coherence tomography angiography may detect distinct vascular features in choroidal nevi and small choroidal melanomas.

Retinal Imaging Using a Confocal Scanning Laser Ophthalmoscope-Based High-Magnification Module.

PURPOSE: To evaluate the usefulness of a high-magnification module (HMM) lens to visualize retinal photoreceptors, retinal nerve fiber layer (RNFL), and superficial retinal vasculature in physiologic and pathologic retinal conditions. DESIGN: Observational descriptive study. PARTICIPANTS: Thirty-two participants with normal and pathologic retina examination results. METHODS: Normal and pathologic maculae were imaged in vivo using still and video HMM lens modes, with fixation and contrast adjustments to enhance visualization. The HMM images were classified qualitatively based on structures identified as either good (photoreceptors seen), average (photoreceptor mosaic cannot be visualized clearly, retinal vessels and other retinal changes can be seen), or poor (no identifiable structures). Selected eyes were imaged with fundus photography, OCT, OCT angiography, indocyanine green angiography, and fluorescein angiography for comparison with the pathologic maculae. MAIN OUTCOMES MEASURES: Description of HMM module-obtained macula images. RESULTS: From 32 eyes imaged (16 normal and 16 pathologic retinas), 12 of 16 normal and 11 of 16 pathologic retinas demonstrated at least average image quality, in which retinal vasculature and landmarks could be visualized. The mosaic pattern of hexagonal shapes representing photoreceptors could not be resolved in most pathologic retinas. For the retinas in which the photoreceptor mosaics were visualized (12 of 16 normal and 2 of 16 pathologic retinas), parafoveal mosaic patterns appeared denser with better image quality for all participants compared with foveal photoreceptors. Difficulty in resolving the photoreceptors in the umbo, fovea, and perifovea was encountered, similar to what has been reported with adaptive optics devices. The RNFL was seen as arcuate hyperreflective bundles. Flow was observed in the macular microvasculature. Poorly resolved photoreceptors and scattered hyperreflective foci were correlated with changes in the retinal pigment epithelium in eyes with age-related macular degeneration or central serous chorioretinopathy. Macular striae were seen in eyes with epiretinal membrane. CONCLUSIONS: In most eyes, regardless of whether retinal pathologic features were present, it was challenging to obtain average quality (or better) images. In the few participants with good-quality imaging, the parafoveal photoreceptor mosaic, vascular flow, and various features of pathologic eyes could be visualized.

Can the Onset of Neovascular Age-Related Macular Degeneration Be an Acceptable Endpoint for Prophylactic Clinical Trials?

Many studies over the past 20 years have pursued the goal of preventing or deferring progression from early and intermediate age-related macular degeneration (AMD) to advanced AMD. The onset of neovascular AMD has been used as a primary endpoint in some prophylactic clinical trials because it is easy to assess and relatively well-defined. Nevertheless, the use of this endpoint for assessing progression of AMD lacks validation. The aims of this paper are to review the current practice of clinical trials investigating the prevention of progression of early or intermediate AMD to neovascular AMD, so-called prophylactic trials, as well as identify ongoing efforts to standardize endpoints and select the ideal population for such studies.

Choroidal nonperfusion on optical coherence tomography angiography in a case of unilateral posterior segment ocular sarcoidosis misdiagnosed as MEWDS.

PURPOSE: To report a case of presumed ocular sarcoidosis initially presenting with features of multiple evanescent white dot syndrome (MEWDS) with atypical optical coherence tomography angiography (OCTA) findings. OBSERVATIONS: A 23 year-old woman presented with a unilateral central scotoma, photophobia, and decreased visual acuity after a viral illness. Examination of the right eye revealed multiple round white macular spots and stippled granularity at the fovea. Multimodal imaging with fluorescein angiography (FA), indocyanine green angiography (ICG), fundus autofluorescence (FAF), and optical coherence tomography (OCT) was consistent with a diagnosis of MEWDS. However, OCTA demonstrated choriocapillaris (CC) flow deficits, which is not typical for MEWDS. The clinical course was initially consistent with MEWDS, with spontaneous recovery of symptoms over ensuing months. The patient presented five months later with floaters and a central scotoma. Examination showed panuveitis, and systemic evaluation revealed an elevated angiotensin converting enzyme (ACE) and hilar lymphadenopathy on chest x-ray consistent with presumed sarcoidosis. CONCLUSIONS AND IMPORTANCE: A case of MEWDS atypically demonstrated CC flow deficits on OCTA and subsequently presented as uveitis secondary to presumed sarcoidosis. Atypical features in MEWDS may be a sign of another disorder masquerading early on as MEWDS and ought to prompt further investigation.

The long-term effects of anti-vascular endothelial growth factor therapy on the optical coherence tomography angiographic appearance of neovascularization in age-related macular degeneration.

BACKGROUND: The short-term effects of anti-vascular endothelial growth factor (anti-VEGF) treatment on macular neovascularization (MNV) morphology is well described, but long-term studies on morphologic changes and correlation of such changes to the type of MNV have not been conducted. This study aims to determine if different types of MNVs in neovascular AMD (nAMD) behave differently with anti-VEGF treatment as visualized on optical coherence tomography angiography (OCTA). METHODS: Treatment-naïve nAMD patients were retrospectively screened for baseline and follow-up OCTA imaging 10 or more months after initial treatment. Images were graded for MNV type, area, activity, mature versus immature vessels, vessel density, presence of atrophy, atrophy location and area. Growth rate was calculated as the percent change in lesion area from baseline over the years of follow-up. In addition, the occurrence of complete regression and the percent of lesions that grew, remained stable, and shrunk per type was also evaluated. RESULTS: Forty-three eyes from 43 patients with a mean follow-up of 2 years were evaluated. On structural OCT, 26 lesions were classified as pure type 1 MNVs, 12 MNVs had a type 2 component, and 5 MNVs had a type 3 component. Of these cases, 2 mixed-type MNVs were considered to have completely regressed. There was no significant differences in MNV area and growth rate between type 1 and type 2 lesions, but all cases of type 3 lesions shrunk in the follow-up period. There was no correlation between the number of injections per year and growth rate, endpoint MNV area or endpoint activity status for any MNV type. There was no significant association between the development of atrophy and the number of injections, baseline MNV area, baseline vessel density, or lesion growth rate. CONCLUSIONS: In nAMD, complete regression of an MNV network exposed to anti-VEGF is rare. This work emphasizes the role of anti-VEGF as anti-leakage rather than vascular regression agents in nAMD.

Model-to-Data Approach for Deep Learning in Optical Coherence Tomography Intraretinal Fluid Segmentation.

Importance: Amid an explosion of interest in deep learning in medicine, including within ophthalmology, concerns regarding data privacy, security, and sharing are of increasing importance. A model-to-data approach, in which the model itself is transferred rather than data, can circumvent many of these challenges but has not been previously demonstrated in ophthalmology. Objective: To determine whether a model-to-data deep learning approach (ie, validation of the algorithm without any data transfer) can be applied in ophthalmology. Design, Setting, and Participants: This single-center cross-sectional study included patients with active exudative age-related macular degeneration undergoing optical coherence tomography (OCT) at the New England Eye Center from August 1, 2018, to February 28, 2019. Data were primarily analyzed from March 1 to June 20, 2019. Main Outcomes and Measures: Training of the deep learning model, using a model-to-data approach, in recognizing intraretinal fluid (IRF) on OCT B-scans. Results: The model was trained (learning curve Dice coefficient, >80%) using 400 OCT B-scans from 128 participants (69 female [54%] and 59 male [46%]; mean [SD] age, 77.5 [9.1] years). In comparing the model with manual human grading of IRF pockets, no statistically significant difference in Dice coefficients or intersection over union scores was found (P > .05). Conclusions and Relevance: A model-to-data approach to deep learning applied in ophthalmology avoided many of the traditional hurdles in large-scale deep learning, including data sharing, security, and privacy concerns. Although the clinical relevance of these results is limited at this time, this proof-of-concept study suggests that such a paradigm should be further examined in larger-scale, multicenter deep learning studies.

Repeatability and Reproducibility of Photoreceptor Density Measurement in the Macula Using the Spectralis High Magnification Module.

PURPOSE: To evaluate the repeatability and reproducibility of photoreceptor density assessment with manual cell counting in healthy participants imaged with the Heidelberg Spectralis High Magnification Module (HMM). DESIGN: Precision study, evaluation of diagnostic test or technology. PARTICIPANTS: Eleven eyes of 8 participants. METHODS: Images were acquired using the Spectralis HMM by a single operator during 2 separate imaging sessions. The 3 highest-quality images of each eye from each session were selected for analysis and coregistered. For a subset of participants, a second operator acquired images in 1 session, and images with the best quality were selected for analysis. Photoreceptor densities were obtained by manual counts in squares of 0.0625 mm2 located in the parafovea. Repeatability (intragrader and intrasession) and reproducibility (interoperator, intergrader, and intersession) were assessed by calculating the intraclass correlation coefficient (ICC) from linear mixed effects models. Bland-Altman plots, coefficients of repeatability, and Pearson correlation results were reported. MAIN OUTCOME MEASURES: Intragrader, intrasession, intersession, interoperator, and intergrader ICC estimates and their 95% confidence intervals for photoreceptor density measurements in the parafovea. RESULTS: Twenty-four eyes of 13 healthy participants were imaged initially. Of these, 11 eyes (45.83%) of 8 participants that had at least 3 acceptable images in each session were included in this study. Mean parafoveal photoreceptor density was 14 988 cells/mm2 (standard deviation, 1403.15 cells/mm2). Intragrader ICC was 0.84 (95% confidence interval, 0.57-0.95), intrasession ICC was 0.69 (95% confidence interval, 0.17-0.86), intersession ICC was 0.88 (95% confidence interval, 0.53-0.96), interoperator ICC was 0.70 (95% confidence interval, 0-0.95), and intergrader ICC was 0.22 (95% confidence interval, 0-0.71). CONCLUSIONS: Images obtained with the HMM allow for photoreceptor mosaic visualization in the macular area, mainly in the parafovea. Although densities obtained are in accordance with other reported methods in the literature, variability within and between images of the apparent cell mosaic were observed, and this study did not demonstrate high repeatability or reproducibility for quantitative assessments using the manual counting method.