Deep-learning based automated quantification of critical optical coherence tomography features in neovascular age-related macular degeneration.

PURPOSE: To validate a deep learning algorithm for automated intraretinal fluid (IRF), subretinal fluid (SRF) and neovascular pigment epithelium detachment (nPED) segmentations in neovascular age-related macular degeneration (nAMD). METHODS: In this IRB-approved study, optical coherence tomography (OCT) data from 50 patients (50 eyes) with exudative nAMD were retrospectively analysed. Two models, A1 and A2, were created based on gradings from two masked readers, R1 and R2. Area under the curve (AUC) values gauged detection performance, and quantification between readers and models was evaluated using Dice and correlation (R2) coefficients. RESULTS: The deep learning-based algorithms had high accuracies for all fluid types between all models and readers: per B-scan IRF AUCs were 0.953, 0.932, 0.990, 0.942 for comparisons A1-R1, A1-R2, A2-R1 and A2-R2, respectively; SRF AUCs were 0.984, 0.974, 0.987, 0.979; and nPED AUCs were 0.963, 0.969, 0.961 and 0.966. Similarly, the R2 coefficients for IRF were 0.973, 0.974, 0.889 and 0.973; SRF were 0.928, 0.964, 0.965 and 0.998; and nPED were 0.908, 0.952, 0.839 and 0.905. The Dice coefficients for IRF averaged 0.702, 0.667, 0.649 and 0.631; for SRF were 0.699, 0.651, 0.692 and 0.701; and for nPED were 0.636, 0.703, 0.719 and 0.775. In an inter-observer comparison between manual readers R1 and R2, the R2 coefficient was 0.968 for IRF, 0.960 for SRF, and 0.906 for nPED, with Dice coefficients of 0.692, 0.660 and 0.784 for the same features. CONCLUSIONS: Our deep learning-based method applied on nAMD can segment critical OCT features with performance akin to manual grading.

BASELINE SATTLER LAYER-CHORIOCAPILLARIS COMPLEX THICKNESS CUTOFFS ASSOCIATED WITH AGE-RELATED MACULAR DEGENERATION PROGRESSION.

PURPOSE: To assess the relationship between choroidal overall and sublayer thickness and age-related macular degeneration (AMD) stage progression. METHODS: A prospective, observational case series was performed. Two hundred and sixty-two eyes of 262 patients with different stages of AMD were imaged by optical coherence tomography. Age-related macular degeneration stage, choroidal thickness, Sattler layer-choriocapillaris complex thickness (SLCCT), and Haller layer thickness were determined at the baseline visit, at a 1-year follow-up visit, at a 2-year follow up visit, and at a final visit (performed after a mean of 5 ± 1 year from the baseline visit). RESULTS: Baseline AMD stages were distributed as follows: early AMD (30 eyes; 12%), intermediate AMD (97 eyes; 39%), and late AMD (126 eyes; 49%). At the final follow-up, AMD stages were so distributed: early AMD (14 eyes; 6%), intermediate AMD (83 eyes; 33%), and late AMD (156 eyes; 61%). Each group showed a statistically significant decrease in choroidal thickness values over the entire follow-up ( P < 0.001), and SLCCT reduction was associated with AMD progression ( P < 0.001). Moreover, SLCCT quantitative cutoffs of <20.50 µ m and <10.5 µ m were associated with a moderate and high probability of AMD progression, respectively, and SLCCT quantitative cutoffs of <18.50 µ m and <8.50 µ m implied a moderate and high probability of macular neovascularization onset, respectively. CONCLUSION: Progressive choroidal impairment contributes to AMD progression. Among choroidal layers, a reduced SLCCT is a promising biomarker of disease worsening, and its quantitative evaluation could help to identify patients at higher risk of stage advancement.

Morphological and Functional Relationship Between OCTA and FA/ICGA Quantitative Features in AMD-Related Macular Neovascularization.

Background: The aim was to study the relationship between quantitative information provided by optical coherence tomography (OCT) angiography (OCTA) and conventional angiography in macular neovascularization (MNV) secondary to age-related macular degeneration (AMD). Methods: The research was designed as an interventional, prospective study. We included 66 eyes (66 patients) affected by naïve MNV. Multimodal imaging included structural OCT, OCTA, fluorescein angiography (FA), and indocyanine green angiography (ICGA). The follow-up lasted 1 year. Patients were treated by PRN anti-VEGF injections. Based on FA/ICGA examinations, we divided the patients into two categories: low vessel tortuosity (VT) (<8.40) and high VT (>8.40), correlating VT with the MNV area, leakage area, speckled fluorescence (SF) quadrants and MNV area/leakage area ratio. Results: Mean baseline BCVA was 0.50 ± 0.61 LogMAR, improved to 0.31 ± 0.29 LogMAR after 1 year (p < 0.01), with a mean number of 7 ± 2 anti-VEGF injections. The patients revealed type-1 MNV in 36 eyes (55%), mixed type 1 and 2 MNV in 18 eyes (27%), and type-2 MNV in 12 eyes (18%). MNV eyes in high-VT MNV featured poorer BCVA, CMT, and OCTA parameters, higher SF quadrants, and less exudation, compared with low-VT MNV (p < 0.01). Moreover, 30% of high-VT MNV eyes developed outer retinal atrophy. Conclusions: Low VT MNV turned out to be more exudative at the baseline but less damaging to the outer retinal structures, whereas high VT MNV proved to be less exudative but more prone to lead to atrophic changes and visual function deterioration. VT may be usefully applied to artificial intelligence-based models designed to characterize MNV secondary to AMD.

Impact of Structural Changes on Multifocal Electroretinography in Patients With Use of Hydroxychloroquine.

Purpose: To investigate the relationship between retinal structure and macular function in eyes screened for hydroxychloroquine (HCQ) toxicity. Methods: Participants referred for hydroxychloroquine retinopathy screening with spectral domain optical coherence tomography (SD-OCT) and multifocal electroretinogram (mfERG) testing were included in the analysis. Amplitude and implicit time of mfERG N1 and P1 responses were included in the analysis. Ring ratios were computed for amplitude values as the ratio of rings 1-3:5 (R1-3:R5). A control group of healthy participants was included for comparison of SD-OCT metrics. Results: Sixty-three eyes screened for HCQ retinopathy and 30 control eyes were analyzed. The outer nuclear layer (ONL) was significantly thinner in HCQ patients in the foveal (P = 0.008), parafoveal (P < 0.0001), and perifoveal (P < 0.0001) regions. The HCQ cohort was further divided into two subgroups according to the presence of structural clinically detectable retinopathy (i.e., structural damage as detected by multimodal imaging). HCQ eyes without retinopathy had a thinner ONL thickness in the foveal (P = 0.032), parafoveal (P < 0.0001), and perifoveal (P < 0.0001) regions and a thinner inner nuclear layer (INL) in the parafoveal region (P = 0.045 versus controls). Structural changes in HCQ patients without retinopathy were significantly associated with macular function as R2:R5 ring ratio of mfERG P1 amplitude was associated with INL (P = 0.002) and ONL (P = 0.044) thicknesses, and R3:R5 ring ratio of P1 amplitude was associated with ONL thickness (P = 0.004). Conclusions: Our results suggest that structural alterations secondary to HCQ toxicity may occur in the absence of clinically detectable retinopathy, and this may reflect in an impaired macular function.

Outer retinal tubulation formation and clinical course of advanced age-related macular degeneration.

Outer retinal tubulations (ORT) are a relatively new finding characterizing outer retinal atrophy. The main aim of the present study was to describe ORT development in advanced age-related macular degeneration (AMD) and to assess its relationship with disease's severity. Patients with advanced AMD characterized either by macular neovascularization or geographic atrophy, showing signs of outer retinal disruption or retinal pigment epithelium atrophy on structural optical coherence tomography (OCT) at the inclusion examination were prospectively recruited. All the patients underwent complete ophthalmologic evaluation, structural OCT scans and fundus autofluorescence imaging. The planned follow-up was of 3-years. Main outcome measures were ORT prevalence, mechanism of ORT formation, mean time needed for complete ORT formation, best-corrected visual acuity (BCVA), definitely decreased autofluorescence (DDAF) area, questionably decreased autofluorescence (QDAF) area, retinal layer thickness, foveal sparing, number of intravitreal injections. We also assessed the possible role of external limiting membrane (ELM) and Müller cells in ORT pathogenesis. Seventy eyes (70 patients) were included; 43 showed dry AMD evolving to geographic atrophy, while 27 displayed the features of wet AMD. Baseline BCVA was 0.5 ± 0.5 LogMAR, decreasing to 0.9 ± 0.5 LogMAR at the 3-year follow-up (p < 0.01). We detected completely formed ORT in 26/70 eyes (37%), subdivided as follows: 20 eyes (77%) wet AMD and 6 eyes (23%) dry AMD (p < 0.01). ORT took 18 ± 8 months (range 3-35 months) to develop fully. We described the steps leading to ORT development, characterized by progressive involvement of, and damage to the photoreceptors, the ELM and the RPE. Eyes displaying ORT were associated with a smaller QDAF area, less retinal layers damage and lower rate of foveal sparing than eyes free of ORT (p < 0.01). We also described pigment accumulations simulating ORT, which were detected in 16/70 eyes (23%), associated with a greater loss of foveal sparing, increased DDAF area and smaller QDAF area at the 3-year follow-up (p < 0.01). In conclusion, this study provided a description of the steps leading to ORT development in AMD. ELM and Müller cells showed a role in ORT pathogenesis. Furthermore, we described a subtype of pigment hypertrophy mimicking ORT, evaluating its clinical utility.

QUANTITATIVE OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY PARAMETER VARIATIONS AFTER TREATMENT OF MACULAR NEOVASCULARIZATION SECONDARY TO AGE-RELATED MACULAR DEGENERATION.

PURPOSE: Macular neovascularization (MNV) secondary to age-related macular degeneration can be characterized by quantitative optical coherence tomography angiography. The aim of the study was to assess the evolution of quantitative optical coherence tomography angiography parameters after 1 year of antivascular endothelial growth factor injections. METHODS: Naive age-related macular degeneration-related MNV eyes were prospectively recruited to analyze optical coherence tomography and optical coherence tomography angiography parameters, including MNV vessel tortuosity (VT) and reflectivity, at baseline and at the end of the follow-up. Macular neovascularization eyes were categorized by a MNV VT cutoff, and quantitative parameter variations were documented after 1 year of treatment. We divided MNV eyes into Group 1 (MNV VT < 8.40) and Group 2 (MNV VT > 8.40). RESULTS: Thrity naive age-related macular degeneration-related MNV eyes (30 patients) were included. Our cohort included 18 Type 1 MNV and 12 Type 2 MNV lesions. Baseline central macular thickness (411 ± 85 µm) improved to 323 ± 54 µm at 1 year (P < 0.01). Only Group 1 MNV displayed significant visual improvement. Macular neovascularization VT values remained stable over the follow-up in both subgroups. Group 2 MNV eyes showed increased MNV reflectivity and increased MNV area at the end of the follow-up. Quantitative retinal capillary plexa parameters were found to be worse in Group 2 MNV. Outer retinal atrophy occurred in 2 of the 18 eyes in MNV Group 1 (11%) and in 6 of the 12 eyes in MNV Group 2 (50%) after 1 year. Vessel density proved to be always worse in Group 2 than in Group 1. CONCLUSION: Macular neovascularization VT provides information on the blood flow and identifies two subgroups with different final anatomical and visual outcomes, regardless of the treatment effect.

Large choroidal excavation in myopic macular degeneration: A case report.

PURPOSE: To report the morphological and clinical features of a case of myopic macular degeneration with large choroidal excavation. METHODS: A myopic patient underwent multimodal imaging, including optical coherence tomography and fluorescein angiography, over a 8-year follow-up. RESULTS: A choroidal excavation was found in left eye, superior to the fovea. The excavation started as a focal choroidal excavation (FCE) and got deeper and larger during the 8-years-long follow-up, thus resulting in a large choroidal excavation (LCE). CONCLUSIONS: LCE may be the evolution of FCE in highly myopic eyes, further studies are needed to describe the natural history of choroidal excavations in degenerative myopia.

Widefield Optical Coherence Tomography Angiography in Diabetic Retinopathy.

PURPOSE: To summarize the role of widefield optical coherence tomography angiography (WF-OCTA) in diabetic retinopathy (DR), extending from the acquisition strategies to the main clinical findings. METHODS: A PubMed-based search was carried out using the terms "Diabetic retinopathy", "optical coherence tomography angiography", "widefield imaging", and "ultra-widefield imaging". All studies published in English up to August 2020 were reviewed. RESULTS: WF-OCTA can be obtained with different approaches, offering advantages over traditional imaging in the study of nonperfusion areas (NPAs) and neovascularization (NV). Quantitative estimates and topographic distribution of NPA and NV are useful for treatment monitoring and artificial intelligence-based approaches. Curvature, segmentation, and motion artifacts should be assessed when using WF-OCTA. CONCLUSIONS: WF-OCTA harbors interesting potential in DR because of its noninvasiveness and capability of objective metrics of retinal vasculature. Further studies will facilitate the migration from traditional imaging to WF-OCTA in both the research and clinical practice fields.