Eduardo Navajas

Segmentation of the foveal microvasculature using deep learning networks

Abstract. Accurate segmentation of the retinal microvasculature is a critical step in the quantitative analysis of the retinal circulation, which can be an important marker in evaluating the severity of retinal diseases. As manual segmentation remains the gold standard for segmentation of optical coherence tomography angiography (OCT-A) images, we present a method for automating the segmentation of OCT-A images using deep neural networks (DNNs). Eighty OCT-A images of the foveal region in 12 eyes from 6 healthy volunteers were acquired using a prototype OCT-A system and subsequently manually segmented. The automated segmentation of the blood vessels in the OCT-A images was then performed by classifying each pixel into vessel or nonvessel class using deep convolutional neural networks. When the automated results were compared against the manual segmentation results, a maximum mean accuracy of 0.83 was obtained. When the automated results were compared with inter and intrarater accuracies, the automated results were shown to be comparable to the human raters suggesting that segmentation using DNNs is comparable to a second manual rater. As manually segmenting the retinal microvasculature is a tedious task, having a reliable automated output such as automated segmentation by DNNs, is an important step in creating an automated output.

Retinal toxicity of high-dose hydroxychloroquine in patients with chronic graft-versus-host disease

OBJECTIVE To evaluate retinal toxicity in patients treated with high-dose hydroxychloroquine (HCQ) (Plaquenil, Sanofi Pharmaceuticals) for chronic graft-versus-host disease (GVHD). DESIGN Cohort study. PARTICIPANTS Twelve patients with chronic GVHD treated with 800 mg/day HCQ between June 2005 and December 2010. METHODS Patients in this study underwent ophthalmologic examination yearly and ancillary studies including colour vision, Amsler grid, fundus photographs, Humphrey 10-2 automated perimetry, spectral-domain optical coherence tomography (SD-OCT), and multifocal electroretinography (mfERG). Evidence of HCQ toxicity was determined by the presence of scotomas in the Amsler grid and Humphrey 10-2 automated perimetry, and confirmed by at least 1 objective test including SD-OCT or mfERG. RESULTS Of the 12 patients, 7 were male and 5 were female. Mean age was 49 years. Mean best corrected visual acuity at baseline was 20/25 and remained 20/25 at final follow-up. Median duration of HCQ treatment was 22.8 months. Median adjusted daily dosage was 11.5 mg/kg/day. Seven patients developed vortex keratopathy. No signs of pigmentary retinopathy or bulls-eye maculopathy were found in any of the patients. Three patients developed retinal toxicity with scotomas in the Amsler grid and Humphrey 10-2 automated perimetry, as well as abnormal mfERG. Retinal structure measured by SD-OCT was abnormal in 2 of the 3 patients with retinal toxicity. Colour vision measured by Ishihara plates, as well as by 100 Hue colour test, was abnormal in 2 of the 3 patients with retinal toxicity. CONCLUSIONS High-dose HCQ in patients with GVHD was associated with higher incidence and earlier development of retinal toxicity.

Strip-based registration of serially acquired optical coherence tomography angiography

Abstract. The visibility of retinal microvasculature in optical coherence tomography angiography (OCT-A) images is negatively affected by the small dimension of the capillaries, pulsatile blood flow, and motion artifacts. Serial acquisition and time-averaging of multiple OCT-A images can enhance the definition of the capillaries and result in repeatable and consistent visualization. We demonstrate an automated method for registration and averaging of serially acquired OCT-A images. Ten OCT-A volumes from six normal control subjects were acquired using our prototype 1060-nm swept source OCT system. The volumes were divided into microsaccade-free en face angiogram strips, which were affine registered using scale-invariant feature transform keypoints, followed by nonrigid registration by pixel-wise local neighborhood matching. The resulting averaged images were presented of all the retinal layers combined, as well as in the superficial and deep plexus layers separately. The contrast-to-noise ratio and signal-to-noise ratio of the angiograms with all retinal layers (reported as average±standard deviation) increased from 0.52±0.22 and 19.58±4.04  dB for a single image to 0.77±0.25 and 25.05±4.73  dB, respectively, for the serially acquired images after registration and averaging. The improved visualization of the capillaries can enable robust quantification and study of minute changes in retinal microvasculature.

Spectral-domain optical coherence tomography in mitochondrial encephalopathy, lactic acidosis, and strokelike episodes.

The point mutation in mitochondrial DNA (mtDNA) 3243A-G is associated with many clinical disorders such as MELAS (mitochondrial encephalomyopathy with lactic acidosis and stroke like episodes) and MIDD (maternally inherited diabetes and deafness). MELAS is characterized by strokelike episodes with headache, episodic nausea and vomiting, and lactic acidosis. Other findings include renal failure, neurosensory hearing loss, diabetes mellitus, and short stature. Ocular findings include pigmentary retinopathy, retinal pigment epithelial (RPE) atrophy, posterior subcapsular cataract, external ophthalmoplegia, and ptosis. In this study, we describe spectral-domain optical coherence tomography (SD-OCT) findings seen in MELAS. A 40-year-old female was referred for evaluation of progressive visual loss and pigmentary retinopathy. Medical history was significant for insulin-dependent diabetes mellitus, neurosensory deafness, short stature, and kidney

Automated identification of cone photoreceptors in adaptive optics optical coherence tomography images using transfer learning

Automated measurements of the human cone mosaic requires the identification of individual cone photoreceptors. The current gold standard, manual labeling, is a tedious process and can not be done in a clinically useful timeframe. As such, we present an automated algorithm for identifying cone photoreceptors in adaptive optics optical coherence tomography (AO-OCT) images. Our approach fine-tunes a pre-trained convolutional neural network originally trained on AO scanning laser ophthalmoscope (AO-SLO) images, to work on previously unseen data from a different imaging modality. On average, the automated method correctly identified 94% of manually labeled cones when compared to manual raters, from twenty different AO-OCT images acquired from five normal subjects. Voronoi analysis confirmed the general hexagonal-packing structure of the cone mosaic as well as the general cone density variability across portions of the retina. The consistency of our measurements demonstrates the high reliability and practical utility of having an automated solution to this problem.

GPU accelerated optical coherence tomography angiography using strip-based registration (Conference Presentation)

High quality visualization of the retinal microvasculature can improve our understanding of the onset and development of retinal vascular diseases, which are a major cause of visual morbidity and are increasing in prevalence. Optical Coherence Tomography Angiography (OCT-A) images are acquired over multiple seconds and are particularly susceptible to motion artifacts, which are more prevalent when imaging patients with pathology whose ability to fixate is limited. The acquisition of multiple OCT-A images sequentially can be performed for the purpose of removing motion artifact and increasing the contrast of the vascular network through averaging. Due to the motion artifacts, a robust registration pipeline is needed before feature preserving image averaging can be performed. In this report, we present a novel method for a GPU-accelerated pipeline for acquisition, processing, segmentation, and registration of multiple, sequentially acquired OCT-A images to correct for the motion artifacts in individual images for the purpose of averaging. High performance computing, blending CPU and GPU, was introduced to accelerate processing in order to provide high quality visualization of the retinal microvasculature and to enable a more accurate quantitative analysis in a clinically useful time frame. Specifically, image discontinuities caused by rapid micro-saccadic movements and image warping due to smoother reflex movements were corrected by strip-wise affine registration estimated using Scale Invariant Feature Transform (SIFT) keypoints and subsequent local similarity-based non-rigid registration. These techniques improve the image quality, increasing the value for clinical diagnosis and increasing the range of patients for whom high quality OCT-A images can be acquired.

Spectral domain optical coherence tomography-guided versus 1-week facedown-posturing after macular hole surgery

OBJECTIVE To compare idiopathic full-thickness macular hole (FTMH) closure rates and postoperative best-corrected visual acuity (BCVA) between spectral domain optical coherence tomography (SD-OCT)-guided facedown posturing and conventional 1-week facedown posturing after macular hole surgery. DESIGN A retrospective comparative study. PARTICIPANTS Thirty-one consecutive eyes that underwent surgery for FTMH between July 2013 and September 2014 were divided into 2 groups. METHODS In the SD-OCT-guided group, SD-OCT was performed on the first day after surgery. If the macular hole was closed, the patient could assume any position but supine. If the hole was not closed on postoperative day 1, SD-OCT was performed daily for 1 week until the hole was closed, at which point the patient was asked to stop positioning. For the control group, patients were asked to keep facedown posturing for 1 week after surgery. RESULTS Mean length of follow-up for the SD-OCT-guided group and the control group was 188.3 (SD = 74.6) days and 216.7 (SD = 71.2) days, respectively. FTMH closure rate was 100% for both groups. Mean time for FTMH closure in the SD-OCT group was 1.2 days. There was no statistical difference in postoperative BCVA at the last visit between SD-OCT-guided (0.49 ± 0.3 logMAR) and control (0.50 ± 0.4 logMAR; p = 0.9422) groups. CONCLUSIONS No difference in FTMH closure rates was detected between the SD-OCT-guided facedown posturing group and the control group. SD-OCT-guided posturing may be used to shorten facedown positioning while maintaining a high success rate.

Acute Macular Neuroretinopathy Associated With Chikungunya Fever.

A 47-year-old man with recent travel to the Caribbean was admitted with acute febrileillness associated with arthralgia and skin rash followed by sudden onset of bilateral visual field defects. Funduscopy revealed subtle bilateral paracentral dark lesions nasal to the fovea best seen on near infrared imaging as hyporeflective, wedge-shaped, paracentral macular lesions. Spectral-domain optical coherence tomography (SD-OCT) through the lesions revealed hyperreflective bands at the level of the outer plexiform layer and outer nuclear layer (ONL), with concomitant attenuation of the underlying external limiting membrane (ELM), ellipsoid zone (EZ), and interdigitation zone (IZ). Neither fluorescein angiography nor speckle variance OCT angiography (sv-OCTA) showed any defects in retinal circulation. Work up revealed positive Immunoglobulin M for Chikungunya virus (CHIKV). Six months later, the patient had persistent scotomas, although reduced in size. SD-OCT showed subtle ONL thinning and restoration of the ELM, although EZ and IZ remained disrupted. Chikungunya fever may manifest as bilateral acute macular neuroretinopathy (AMN). Clinicians should be aware of possible systemic associations of AMN. [Ophthalmic Surg Lasers Imaging Retina. 2016;47:596-599.].