Optical coherence tomography angiography and Humphrey visual field in patients with obstructive sleep apnea.

STUDY OBJECTIVES: To determine if obstructive sleep apnea syndrome (OSAS) predisposes patients to glaucoma and macular disease due to vascular compromise by evaluating retinal and optic nerve vasculature and function using optical coherence tomography angiography and Humphrey visual field testing, respectively. METHODS: In this prospective, observational, cross-sectional study 45 patients undergoing polysomnography ordered per standard of care were selected and stratified based on apnea-hypopnea index (AHI). Medical history, visual acuity testing, 24-2 Humphrey visual field, intraocular pressure measurement, and optical coherence tomography angiography studies of the macular and peripapillary retina were obtained. Correlations between polysomnography parameters and imaging data were analyzed. RESULTS: The radial peripapillary capillary vascular density demonstrated no relationship to AHI (95% confidence interval [CI] [-0.026,0.038]) or severity of OSAS (95% CI: [-0.772, 3.648]) for moderate OSAS compared to mild/normal and (-1.295, 3.1421) for severe compared to mild/normal. Optical coherence tomography angiography superficial parafoveal vascular density (95% CI: [-0.068,0.011], deep parafoveal vascular density (95% CI: [-0.080,0.009]), and foveal avascular zone (95% CI: [-0.001, 0.001]) showed no statistically significant relationship to AHI or OSAS severity after controlling for confounders. Optical coherence tomography retinal nerve fiber layer thickness increased with AHI (P = .014), but there was no statistically significant correlation with OSAS severity with retinal nerve fiber layer thickness (95% CI: [-12.543, 6.792] for moderate comparing to normal and [-2.883, 16.551] for severe comparing to normal). Visual field parameters were unaffected by OSAS (95% CI: mean deviation [-0.21,0.29], pattern standard deviation: [-0.351, 0.121], visual field index: [-0.166, 0.329]). Optical coherence tomography choroidal thickness showed a statistically significant decrease when OSAS was grouped by severity (P = .0092) but did not correlate with AHI (P = .129, 95% CI: [-1.210, 0.095]). CONCLUSIONS: The severity of OSAS did not show a statistically significant effect on parameters associated with glaucoma or macular vascular disease. Larger cohorts may be required to determine the physiologic consequences of OSAS on the macular and optic nerve vasculature, structure, and function. CITATION: Davanian A, Williamson L, Taylor C, et al. Optical coherence tomography angiography and Humphrey visual field in patients with obstructive sleep apnea. J Clin Sleep Med 2022;18(9):2133-2142.

Bilateral Acute Angle Closure in a Pediatric Patient Taking Lisdexamfetamine Dimesylate (Vyvanse).

Attention-deficit/hyperactivity disorder is commonly treated with amphetamines as first line therapy. Rare case reports have shown amphetamines are associated with open angle glaucoma. We report a rare case of a 14-year-old male who presented with bilateral acute angle closure presumed to be related to his use of lisdexamfetamine dimesylate (Vyvanse). The patient's medication was discontinued which resulted in complete resolution of angle closure.

Non-transfer Deep Learning of Optical Coherence Tomography for Post-hoc Explanation of Macular Disease Classification.

Deep transfer learning is a popular choice for classifying monochromatic medical images using models that are pretrained by natural images with color channels. This choice may introduce unnecessarily redundant model complexity that can limit explanations of such model behavior and outcomes in the context of medical imaging. To investigate this hypothesis, we develop a configurable deep convolutional neural network (CNN) to classify four macular disease conditions using retinal optical coherence tomography (OCT) images. Our proposed non-transfer deep CNN model (acc: 97.9%) outperforms existing transfer learning models such as ResNet-50 (acc: 89.0%), ResNet-101 (acc: 96.7%), VGG-19 (acc: 93.3%), Inception-V3 (acc: 95.8%) in the same retinal OCT image classification task. We perform post-hoc analysis of the trained model and model extracted image features, which reveals that only eight out of 256 filter kernels are active at our final convolutional layer. The convolutional responses of these selective eight filters yield image features that efficiently separate four macular disease classes even when projected onto two-dimensional principal component space. Our findings suggest that many deep learning parameters and their computations are redundant and expensive for retinal OCT image classification, which are expected to be more intense when using transfer learning. Additionally, we provide clinical interpretations of our misclassified test images identifying manifest artifacts, shadowing of useful texture, false texture representing fluids, and other confounding factors. These clinical explanations along with model optimization via kernel selection can improve the classification accuracy, computational costs, and explainability of model outcomes.

Mask-induced Artifact Impacts Intraocular Pressure Measurement Using Goldmann Applanation Tonometry.

PURPOSE: The coronavirus (COVID-19) pandemic has impacted ophthalmology practices significantly. American Academy of Ophthalmology and Center for Disease Control guidelines suggest mandatory masking of patients and physicians during outpatient visits. We have recently become aware of a mask-induced phenomenon, whereby the intraocular pressure (IOP) as measured by Goldmann applanation tonometry (GAT) is artificially elevated due to mechanical interference from the mask. CLINICAL PRESENTATION: A 37-year-old male with a history of primary open-angle glaucoma on triple therapy presented for a routine visit. CLINICAL FINDINGS: When measuring IOP by GAT the right eye measured 16 mm Hg, but the left eye measured 20 mm Hg. The patient's mask was noted to be touching the base of the sensor rod on the tonometer. This patient's IOP was falsely elevated due to the lateral edge of his mask touching the base of the applanation tonometer, changing the relationship between the bi-prism tip and the weighted balance below, and eliminating the weighted balance from the pressure measuring mechanism. The patient's mask was adjusted to ensure there was no touch and repeat measurement showed an IOP of 16 mm Hg in the left eye. CONCLUSION: Recognizing mask-induced alteration in IOP is essential as it could lead to unnecessary escalation of treatment. We recommend flattening the area of mask protrusion during applanation and ensuring that the sensor arm remains clear of the mask while the tonometer tip approaches the cornea, especially at the moment the mires become visible during corneal contact.