Deep Learning-Based Automated Detection of Retinal Breaks and Detachments on Fundus Photography.

Purpose: The purpose of this study was to develop a deep learning algorithm, to detect retinal breaks and retinal detachments on ultra-widefield fundus (UWF) optos images using artificial intelligence (AI). Methods: Optomap UWF images of the database were annotated to four groups by two retina specialists: (1) retinal breaks without detachment, (2) retinal breaks with retinal detachment, (3) retinal detachment without visible retinal breaks, and (4) a combination of groups 1 to 3. The fundus image data set was split into a training set and an independent test set following an 80% to 20% ratio. Image preprocessing methods were applied. An EfficientNet classification model was trained with the training set and evaluated with the test set. Results: A total of 2489 UWF images were included into the dataset, resulting in a training set size of 2008 UWF images and a test set size of 481 images. The classification models achieved an area under the receiver operating characteristic curve (AUC) on the testing set of 0.975 regarding lesion detection, an AUC of 0.972 for retinal detachment and an AUC of 0.913 for retinal breaks. Conclusions: A deep learning system to detect retinal breaks and retinal detachment using UWF images is feasible and has a good specificity. This is relevant for clinical routine as there can be a high rate of missed breaks in clinics. Future clinical studies will be necessary to evaluate the cost-effectiveness of applying such an algorithm as an automated auxiliary tool in a large practices or tertiary referral centers. Translational Relevance: This study demonstrates the relevance of applying AI in diagnosing peripheral retinal breaks in clinical routine in UWF fundus images.

High-Resolution Optical Coherence Tomography in Healthy Individuals Provides Resolution at the Cellular and Subcellular Levels.

Purpose: To assess the clinical resolution capacities of a novel high-resolution optical coherence tomography (High-Res OCT). Methods: Eight healthy volunteers were included in this observational study. Using the SPECTRALIS High-Res OCT device (Heidelberg Engineering, Heidelberg) macular b-scans were taken and compared with b-scans acquired with a SPECTRALIS HRA+OCT device (Heidelberg Engineering, Heidelberg). High-Res OCT scans were also compared with hematoxylin and eosin-stained sections from a human donor retina. Results: High-Res OCT allowed identification of several retinal structures at the cellular and subcellular levels, namely, cell nuclei of ganglion cells, displaced amacrine cells, cone photoreceptors and retinal pigment epithelial cells compared with the commercial device. Rod photoreceptor nuclei were partially detectable. Localization of cell type-specific nuclei were confirmed by histological sections of human donor retina. Additionally, all three plexus of the retinal vasculature could be visualized. Conclusions: SPECTRALIS High-Res OCT device provides improved resolution compared with the conventional SPECTRALIS HRA+OCT device and allows to identify structures at the cellular level, similar to histological sections. Translational Relevance: High-Res OCT shows improved visualization of retinal structures in healthy individuals and can be used to assess individual cells within the retina.

Fluorescence lifetime imaging ophthalmoscopy and the influence of oral lutein/zeaxanthin supplementation on macular pigment (FLOS) - A pilot study.

BACKGROUND & AIMS: Oral lutein (L) and zeaxanthin (Z) supplementation enhances macular pigment optical density (MPOD) and plays a protective role in the development of age-related macular degeneration (AMD). Fluorescence lifetime imaging ophthalmoscopy (FLIO) is a novel in vivo retinal imaging method that has been shown to correlate to classical MPOD measurements and might contribute to a metabolic mapping of the retina in the future. Our aim was to show that oral supplementation of L and Z affects the FLIO signal in a positive way in patients with AMD. METHODS: This was a prospective, single center, open label cohort study. Patients with early and intermediate AMD received oral L and Z supplementation during three months, and were observed for another three months after therapy termination. All visits included measurements of clinical parameters, serum L and Z concentration, MPOD measurements using heterochromatic flicker photometry, dual wavelength autofluorescence imaging, and FLIO. Correlation analysis between FLIO and MPOD were performed. RESULTS: Twenty-one patients completed the follow up period. Serum L and Z concentrations significantly increased during supplementation (mean difference 244.8 ng/ml; 95% CI: 81.26-419.9, and 77.1 ng/ml; 95% CI: 5.3-52.0, respectively). Mean MPOD units significantly increased (mean difference 0.06; 95% CI: 0.02-0.09; at 0.5°, 202; 95% CI: 58-345; at 2°, 1033; 95% CI: 288-1668; at 9° of eccentricity, respectively) after three months of supplementation with macular xanthophylls, which included L and Z. Median FLIO lifetimes in the foveal center significantly decreased from 277.3 ps (interquartile range 230.2-339.1) to 261.0 ps (interquartile range 231.4-334.4, p = 0.027). All parameters returned to near-normal values after termination of the nutritional supplementation. A significant negative correlation was found between FLIO and MPOD (r2 = 0.57, p < 0.0001). CONCLUSIONS: FLIO is able to detect subtle changes in MPOD after L and Z supplementation in patients with early and intermediate AMD. Our findings confirm the previous described negative correlation between FLIO and MPOD. Macular xanthophylls seem to contribute to short foveal lifetimes. This study is registered at ClinicalTrials.gov (identifier number NCT04761341).

FLUORESCENCE LIFETIME IMAGING OPHTHALMOSCOPY AS PREDICTOR OF LONG-TERM FUNCTIONAL OUTCOME IN MACULA-OFF RHEGMATOGENOUS RETINAL DETACHMENT.

PURPOSE: To assess whether macular fluorescence lifetimes may serve as a predictor for long-term outcomes in macula-off rhegmatogenous retinal detachment. METHODS: A single-center observational study was conducted. Patients with pseudophakic macula-off rhegmatogenous retinal detachment were included and evaluated 1 and 6 months after successful reattachment surgery. Fluorescence lifetime imaging ophthalmoscopy lifetimes in the central Early Treatment Diabetic Retinopathy Study grid subfield, in two distinct channels (short spectral channel and long spectral channel) were analyzed. Best-corrected visual acuity optical coherence tomography of the macula and fluorescence lifetimes were measured at month 1 and month 6. RESULTS: Nineteen patients were analyzed. Lifetimes of the previously detached retinas were prolonged compared with the healthy fellow eyes. Short lifetimes at month 1 were associated with better best-corrected visual acuity improvement (short spectral channel: r2 = 0.27, P < 0.05, long spectral channel: r2 = 0.23, P < 0.05) and with good final best-corrected visual acuity (short spectral channel: r2 = 0.43, P < 0.01, long spectral channel: r2 = 0.25, P < 0.05). Lifetimes were prolonged in some cases of outer retinal damage in optical coherence tomography scans. CONCLUSION: Fluorescence lifetime imaging ophthalmoscopy might serve as a prediction tool for functional recovery in pseudophakic macula-off rhegmatogenous retinal detachment. Retinal fluorescence lifetimes could give insight in molecular processes after rhegmatogenous retinal detachment.

Evaluation of an Artificial Intelligence-Based Detector of Sub- and Intraretinal Fluid on a Large Set of Optical Coherence Tomography Volumes in Age-Related Macular Degeneration and Diabetic Macular Edema.

INTRODUCTION: In this retrospective cohort study, we wanted to evaluate the performance and analyze the insights of an artificial intelligence (AI) algorithm in detecting retinal fluid in spectral-domain OCT volume scans from a large cohort of patients with neovascular age-related macular degeneration (AMD) and diabetic macular edema (DME). METHODS: A total of 3,981 OCT volumes from 374 patients with AMD and 11,501 OCT volumes from 811 patients with DME were acquired with Heidelberg-Spectralis OCT device (Heidelberg Engineering Inc., Heidelberg, Germany) between 2013 and 2021. Each OCT volume was annotated for the presence or absence of intraretinal fluid (IRF) and subretinal fluid (SRF) by masked reading center graders (ground truth). The performance of an already published AI algorithm to detect IRF and SRF separately, and a combined fluid detector (IRF and/or SRF) of the same OCT volumes was evaluated. An analysis of the sources of disagreement between annotation and prediction and their relationship to central retinal thickness was performed. We computed the mean areas under the curves (AUC) and under the precision-recall curves (AP), accuracy, sensitivity, specificity, and precision. RESULTS: The AUC for IRF was 0.92 and 0.98, for SRF 0.98 and 0.99, in the AMD and DME cohort, respectively. The AP for IRF was 0.89 and 1.00, for SRF 0.97 and 0.93, in the AMD and DME cohort, respectively. The accuracy, specificity, and sensitivity for IRF were 0.87, 0.88, 0.84, and 0.93, 0.95, 0.93, and for SRF 0.93, 0.93, 0.93, and 0.95, 0.95, 0.95 in the AMD and DME cohort, respectively. For detecting any fluid, the AUC was 0.95 and 0.98, and the accuracy, specificity, and sensitivity were 0.89, 0.93, and 0.90 and 0.95, 0.88, and 0.93, in the AMD and DME cohort, respectively. False positives were present when retinal shadow artifacts and strong retinal deformation were present. False negatives were due to small hyporeflective areas in combination with poor image quality. The combined detector correctly predicted more OCT volumes than the single detectors for IRF and SRF, 89.0% versus 81.6% in the AMD and 93.1% versus 88.6% in the DME cohort. DISCUSSION/CONCLUSION: The AI-based fluid detector achieves high performance for retinal fluid detection in a very large dataset dedicated to AMD and DME. Combining single detectors provides better fluid detection accuracy than considering the single detectors separately. The observed independence of the single detectors ensures that the detectors learned features particular to IRF and SRF.

Prevalence and Mimics of Kleine-Levin Syndrome: A Survey in French-Speaking Switzerland.

STUDY OBJECTIVE: Kleine-Levin syndrome (KLS) is a rare disease of unknown etiology, the diagnosis of which can be challenging. We aimed to estimate KLS prevalence in French-speaking Switzerland, and assess differences with mimicking conditions. METHODS: In this cross-sectional study, KLS patients were identified through a population-based approach, including at our hospital and contacting all sleep-certified facilities and neurologists in French-speaking Switzerland. Furthermore, we identified patients referred to our center for suspected KLS that received other diagnoses. Relevant clinical data of these two groups was compared. RESULTS: We identified 7 patients with diagnosed KLS (6 since 2009), leading to a prevalence estimation of 3.19 per million (95% confidence interval: 1.55-6.59). Median age at diagnosis was 17 years (range: 12-19), 71.4% of them were men, and mean diagnosis delay after the first episode was 20.1 ± 10.9 months. We identified 9 mimic patients referred to our center; they differed from KLS patients by their higher age at disease onset (median: 15 [range: 12-16] vs. 19 [range: 16-64] years; p < 0.001), suspected KLS as referral reason (more frequent in mimics, p = 0.003), and the presence of precipitating factors (more frequent in KLS, p = 0.011). Among the mimic patients, 77% (versus 28% in KLS) had a psychiatric diagnosis. CONCLUSIONS: This study suggests a relatively higher KLS prevalence than previously reported. As compared to KLS, mimic patients have higher age at symptom onset, are more often initially referred for KLS suspicion, and have a higher prevalence of psychiatric disorders.