Search

Purpose: To develop machine learning (ML) models to predict, at baseline, treatment outcomes at month 9 in patients with neovascular age-related macular degeneration (nAMD) receiving faricimab. Design: Retrospective proof of concept study. Participants: Patients enrolled in the phase II AVENUE trial (NCT02484690) of faricimab in nAMD. Methods: Baseline characteristics and spectral domain-OCT (SD-OCT) image data from 185 faricimab-treated eyes were split into 80% training and 20% test sets at the patient level. Input variables were baseline age, sex, best-corrected visual acuity (BCVA), central subfield thickness (CST), low luminance deficit, treatment arm, and SD-OCT images. A regression problem (BCVA) and a binary classification problem (reduction of CST by 35%) were considered. Overall, 10 models were developed and tested for each problem. Benchmark classical ML models (linear, random forest, extreme gradient boosting) were trained on baseline characteristics; benchmark deep neural networks (DNNs) were trained on baseline SD-OCT B-scans. Baseline characteristics and SD-OCT data were merged using 2 approaches: model stacking (using DNN prediction as an input feature for classical ML models) and model averaging (which averaged predictions from the DNN using SD-OCT volume and from classical ML models using baseline characteristics). Main Outcome Measures: Treatment outcomes were defined by 2 target variables: functional (BCVA letter score) and anatomical (percent decrease in CST from baseline) outcomes at month 9. Results: The best-performing BCVA regression model with respect to the test coefficient of determination (R2) was the linear model in the model-stacking approach with R2 of 0.31. The best-performing CST classification model with respect to test area under receiver operating characteristics (AUROC) was the benchmark linear model with AUROC of 0.87. A post hoc analysis showed the baseline BCVA and the baseline CST had the most effect in the all-model prediction for BCVA regression and CST classification, respectively. Conclusions: Promising signals for predicting treatment outcomes from baseline characteristics were detected; however, the predictive benefit of baseline images was unclear in this proof-of-concept study. Further testing and validation with larger, independent datasets is required to fully explore the predictive capacity of ML models using baseline imaging data. Financial Disclosure(s): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Purpose: To assess different microperimetry (MP) macular sensitivity outcome measures capturing functional deterioration in eyes with geographic atrophy (GA) secondary to age-related macular degeneration (AMD). Design: Patients were included from 2 identically designed, phase III, double-masked, randomized controlled clinical trials, Chroma (NCT02247479) and Spectri (NCT02247531). Participants: Patients enrolled were aged ≥ 50 years with bilateral GA and no evidence of previous or active neovascular AMD. Methods: Patients were randomized 2:1:2:1 to receive through 96 weeks intravitreal lampalizumab 10 mg every 4 weeks (LQ4W), every 6 weeks (LQ6W), or corresponding sham procedures. For this study, mesopic macular sensitivity of the central 20° was assessed using MP-1 microperimeter at selected sites. Main Outcome Measures: Two exploratory endpoints were developed, namely perilesional sensitivity (average of points adjacent to absolute scotomatous points) and responding sensitivity (average of all nonscotomatous points; > 0 dB at baseline) by using customized masks for each patient. These were compared with conventional MP endpoints (mean macular sensitivity and number of absolute scotomatous points). Results: Of 1881 Chroma and Spectri participants, 277 agreed to participate in the present study. Of these, 197 (LQ4W, n = 63; LQ6W, n = 68; pooled sham, n = 66) had reliable MP results. Enlargement of GA lesion area by approximately 2 mm2/year across treatment groups was accompanied by deterioration in all MP parameters. There was no difference in worsening of macular sensitivity or absolute scotomatous points among treatment groups. Perilesional and responding sensitivities showed greater decline over time than mean macular sensitivity. Change in GA lesion area at week 48 showed better correlation with perilesional sensitivity (r = −0.17) and responding sensitivity (r = −0.20) than mean macular sensitivity (r = −0.03), while the correlation was highest with the number of absolute scotomatous points (r = 0.37). Conclusions: Perilesional or responding macular sensitivity measured by MP should be considered more sensitive endpoints than mean macular sensitivity for monitoring functional decline over time in GA. Although perilesional, responding, and mean macular sensitivity had weak correlation with GA lesion area, the number of absolute scotomatous points may provide additional information on the anatomic/functional correlation. Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references.

Purpose: Neovascular age-related macular degeneration (nAMD) shows variable treatment response to intravitreal anti-VEGF. This analysis compared the potential of different artificial intelligence (AI)-based machine learning models using OCT and clinical variables to accurately predict at baseline the best-corrected visual acuity (BCVA) at 9 months in response to ranibizumab in patients with nAMD. Design: Retrospective analysis. Participants: Baseline and imaging data from patients with subfoveal choroidal neovascularization secondary to age-related macular dengeration. Methods: Baseline data from 502 study eyes from the HARBOR (NCT00891735) prospective clinical trial (monthly ranibizumab 0.5 and 2.0 mg arms) were pooled; 432 baseline OCT volume scans were included in the analysis. Seven models, based on baseline quantitative OCT features (Least absolute shrinkage and selection operator [Lasso] OCT minimum [min], Lasso OCT 1 standard error [SE]); on quantitative OCT features and clinical variables at baseline (Lasso min, Lasso 1SE, CatBoost, RF [random forest]); or on baseline OCT images only (deep learning [DL] model), were systematically compared with a benchmark linear model of baseline age and BCVA. Quantitative OCT features were derived by a DL segmentation model on the volume images, including retinal layer volumes and thicknesses, and retinal fluid biomarkers, including statistics on fluid volume and distribution. Main Outcome Measures: Prognostic ability of the models was evaluated using coefficient of determination (R2) and median absolute error (MAE; letters). Results: In the first cross-validation split, mean R2 (MAE) of the Lasso min, Lasso 1SE, CatBoost, and RF models was 0.46 (7.87), 0.42 (8.43), 0.45 (7.75), and 0.43 (7.60), respectively. These models ranked higher than or similar to the benchmark model (mean R2, 0.41; mean MAE, 8.20 letters) and better than OCT-only models (mean R2: Lasso OCT min, 0.20; Lasso OCT 1SE, 0.16; DL, 0.34). The Lasso min model was selected for detailed analysis; mean R2 (MAE) of the Lasso min and benchmark models for 1000 repeated cross-validation splits were 0.46 (7.7) and 0.42 (8.0), respectively. Conclusions: Machine learning models based on AI-segmented OCT features and clinical variables at baseline may predict future response to ranibizumab treatment in patients with nAMD. However, further developments will be needed to realize the clinical utility of such AI-based tools. Financial Disclosure(s): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Purpose: To investigate intraretinal neovascularization and microvascular anomalies by correlating in vivo multimodal imaging with corresponding ex vivo histology in a single patient. Design: A case study comprising clinical imaging from a community-based practice, and histologic analysis at a university-based research laboratory (clinicopathologic correlation). Participants: A White woman in her 90s treated with numerous intravitreal anti-VEGF injections for bilateral type 3 macular neovascularization (MNV) secondary to age-related macular degeneration (AMD). Methods: Clinical imaging comprised serial infrared reflectance, eye-tracked spectral-domain OCT, OCT angiography, and fluorescein angiography. Eye tracking, applied to the 2 preserved donor eyes, enabled the correlation of clinical imaging signatures with high-resolution histology and transmission electron microscopy. Main Outcome Measures: Histologic/ultrastructural descriptions and diameters of vessels seen in clinical imaging. Results: Six vascular lesions were histologically confirmed (type 3 MNV, n = 3; deep retinal age-related microvascular anomalies [DRAMAs], n = 3). Pyramidal (n = 2) or tangled (n = 1) morphologies of type 3 MNV originated at the deep capillary plexus (DCP) and extended posteriorly to approach without penetrating persistent basal laminar deposit. They did not enter the subretinal pigment epithelium (RPE)–basal laminar space or cross the Bruch membrane. Choroidal contributions were not found. The neovascular complexes included pericytes and nonfenestrated endothelial cells, within a collagenous sheath covered by dysmorphic RPE cells. Deep retinal age-related microvascular anomaly lesions extended posteriorly from the DCP into the Henle fiber and the outer nuclear layers without evidence of atrophy, exudation, or anti-VEGF responsiveness. Two DRAMAs lacked collagenous sheaths. External and internal diameters of type 3 MNV and DRAMA vessels were larger than comparison vessels in the index eyes and in aged normal and intermediate AMD eyes. Conclusions: Type 3 MNV vessels reflect specializations of source capillaries and persist during anti-VEGF therapy. The collagenous sheath of type 3 MNV lesions may provide structural stabilization. If so, vascular characteristics may be useful in disease monitoring in addition to fluid and flow signal detection. Further investigation with longitudinal imaging before exudation onset will help determine if DRAMAs are part of the type 3 MNV progression sequence. Financial Disclosure(s): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Abstract A paracentesis prior to an intravitreal injection is a very safe procedure and can prevent IOP-spikes after injections. As these spikes pose the risk of inducing glaucomatous changes particularly in patients with frequent injections and/or with a risk profile, a regular paracentesis prior to an injection may be considered and discussed with the patient.

Abstract Introduction The specific impact from the patient’s perspective of geographic atrophy (GA), an advanced form of age-related macular degeneration (AMD), is not well understood. Methods An ethnographic study was conducted to understand the impact of bilateral GA secondary to AMD on daily functioning by observing regular activities performed at home and through semi-structured interviews. Eligible subjects had a definitive GA diagnosis, including presence of drusen, GA lesion size of at least one disc area in the better-seeing eye, and no other confounding ophthalmologic diagnosis. Data were collected via video recordings and field notes, and analyzed by coding video transcripts. Results Functional impact domains affecting more than two of the 16 subjects from the United Kingdom, United States, or Germany were activities of daily living (difficulty reading, n = 16; driving, n = 12; and watching movies, television, or theater, n = 11), emotional (frustration, and fear of blindness, n = 7 each), social/leisure (interference with hobbies, n = 8, and diminished social activities, n = 4), physical (n = 4), and financial (n = 10). Subjects with a best-corrected visual acuity (BCVA) of 20/100 or better in the better-seeing eye (n = 10) reported similar functional impacts to those with a BCVA of worse than 20/100 in their better-seeing eye (n = 5). Conclusion This study helps address gaps in patient-focused research into GA, which negatively impacts the day-to-day functioning of patients. Larger qualitative and quantitative studies are needed to quantify patient experiences and assess the correlation between BCVA score and impact of GA. Funding F. Hoffmann-La Roche Ltd.

Abstract Purpose To analyse morphologic features of the choroid in Non-pathological myopic eyes using spectral-domain (SD) optical coherence tomography (OCT). Methods Retrospective analysis of enhanced depth SD-OCT images of Non-pathological myopic eyes in comparison with age-matched healthy controls was performed. Choroidal thickness (CT) and large choroidal vessel thickness (LCVT) were measured at the fovea, 750 µm nasally from fovea (N750) and 750 µm temporally (T750) from fovea. Medium choroidal vessel thickness (MCVT) was calculated by subtracting LCVT from CT. Choriocapillaris was encompassed by MCVT, given its reduced thickness. Linear regression analysis evaluated the relationship between age and axial with CT, LCVT and MCVT. Results The study group comprised 42 eyes of 31 patients (mean age 46.13 ± 15.63; 15 females). Control group included 57 eyes of 34 patients (mean age of 42.3 ± 15.29; 24 females). Mean axial length in myopic eyes and control group was 26.57 ± 1.27 and 23.59 ± 0.99 mm respectively. Myopic eyes showed significant thinning of MCVT and CT at all locations (p 0.05). With each decade, thinning of up to 37 µm in CT was noted along with thinning of LCVT (up to 22.6 µm) and MCVT (up to 25 µm). Each mm increase in axial length caused 38.2 µm thinning of choroid along with LCVT (30 µm). Conclusion Significant thinning of MCVT was noted in non-pathological myopic eyes in comparison to healthy subjects. It appears that MCVT has stronger relationship with age and axial length.

The purpose of this review is to summarize the current knowledge on choriocapillaris loss in advanced age macular degeneration (AMD). Several histopathological studies in animal models and human eyes had showed that the choriocapillaris density decreases with age. However, the role of choriocapillaris loss is still unclear in AMD and its advanced forms, either choroidal neovascularization (CNV) or geographic atrophy (GA). Some authors have hypothesized that choriocapillaris loss might precede overt retinal pigment epithelium atrophy. Others have hypothesized that deposition of complement complexes on and around the choriocapillaris could be related to the tissue loss observed in early AMD. The development of imaging modalities, such as optical coherence tomography angiography (OCTA), have led to a better understanding of underlying physiopathological mechanisms in AMD. OCTA showed atrophy of choriocapillaris underneath and beyond the region of photoreceptors and RPE loss, in agreement with previous histopathologic studies. The evolution of OCTA technology suggests that CNV seems to originate from regions of severe choriocapillaris alteration. Significant progress has been made in the understanding of development and progression of GA and CNV. In vivo investigation of the choriocapillaris using OCTA may lead to new insights related to underlying disease mechanisms in AMD.

To characterize qualitative and quantitative features of the choroid in normal eyes using enface swept-source optical coherence tomography (SS-OCT).Fifty-two eyes of 26 consecutive normal subjects were prospectively recruited to obtain multiple three-dimensional 12 x 12 mm volumetric scans using a long-wavelength high-speed SS-OCT prototype. A motion-correction algorithm merged multiple SS-OCT volumes to improve signal. Retinal pigment epithelium (RPE) was segmented as the reference and enface images were extracted at varying depths every 4.13 μm intervals. Systematic analysis of the choroid at different depths was performed to qualitatively assess the morphology of the choroid and quantify the absolute thicknesses as well as the relative thicknesses of the choroidal vascular layers including the choroidal microvasculature (choriocapillaris, terminal arterioles and venules; CC) and choroidal vessels (CV) with respect to the subfoveal total choroidal thickness (TC). Subjects were divided into two age groups: younger (

A new gas chromatographic/mass spectrometric (GC/MS) method was developed to detect 2-methylanthraquinone (2-MeA) in wood dust. 2-MeA is present in teak wood (a suspected human carcinogen) but not in oak, beech, mahogany, birch, ash or pine. The method involved collection of workplace dust on filters and extraction of 2-MeA with methanol and GC/MS analysis. The method was tested on teak wood dust samples (n = 43) collected on polyvinylchloride membrane filters during various work operations in four small factories making furniture and fittings for leisure craft and boatyards (air teak wood dust concentration: range 0.32–14.32 mg m−3). A high correlation coefficient for the content of 2-MeA versus teak dust was obtained (logarithmic correlation: y = 1.5308x + 0.0998, r = 0.9215). Determination of airborne 2-MeA is a useful technique to confirm occupational exposure to teak wood dust. [ABSTRACT FROM AUTHOR]