Color Fundus Images Research Articles

An Automated Early Detection of Glaucoma using Support Vector Machine Based Visual Geometry Group 19 (VGG-19) Convolutional Neural Network

Deep learning is a useful technique for investigating the medicinal images. Glaucoma is a neurotic condition, dynamic neuro degeneration of the optic nerve, which leads visual impairment. It could be forestalled by an early detection of glaucoma and the regular screening with specialist for glaucoma diagnosis. Glaucoma is assessed by observing intra ocular pressure and optic Cup-Disc-Ratio (CDR). In this paper, novel mechanized glaucoma recognition has been performed by utilizing computer supported analysis from fundus images. The simulation outcomes are acquired by utilizing a Support Vector Machine based VGG-19 network architecture. The CDR threshold value of 0.41 has been used for glaucoma recognition. The fundus images which has the CDR greater than 0.41 is treated as glaucoma affected and less than 0.41 is non-glaucoma fundus images. The proposed glaucoma recognition system works with reasonable to obtain and generally utilized digital color fundus images. For the set of 175 fundus images a classification precision of 94% has been accomplished.

Automatic Diabetic Retinopathy Diagnosis Using Adaptive Fine-Tuned Convolutional Neural Network

Diabetic retinopathy (DR) is a complication of diabetes that leads to blindness. The manual screening of color fundus images to detect DR at early stages is expensive and time consuming. Deep learning (DL) techniques have been employed for automatic DR screening on fundus images due to their outstanding performance in many applications. However, training a DL model needs a huge amount of data, which are usually unavailable in the case of DR, and overfitting is unavoidable. Employing a two-stage transfer learning method, we developed herein an intelligent computer-aided system using a pre-trained convolutional neural network (CNN) for automatic DR screening on fundus images. A CNN model learns the domain-specific hierarchy of low- to high-level features. Given this, using the regions of interest (ROIs) of lesions extracted from the annotated fundus images, the first layer of a pre-trained CNN model is re-initialized. The model is then fine-tuned, such that the low-level layers learn the local structures of the lesion and normal regions. As the fully connected layer (FC) layers encode high-level features, which are global in nature and domain specific, we replace them with a new FC layer based on the principal component analysis PCA and use it in an unsupervised manner to extract discriminate features from the fundus images. This step reduces the model complexity, significantly avoiding the overfitting problem. This step also lets the model adopt the fundus image structures, making it suitable for DR feature detection. Finally, we add a gradient boosting-based classification layer. The evaluation of the proposed system using a 10-fold cross-validation on two challenging datasets (i.e., EyePACS and Messidor) indicates that it outperforms state-of-the-art methods. It will be useful for the initial screening of DR patients and will help graders in deciding quickly as regards patient referral to an ophthalmologist for further diagnosis and treatment.

Robust Content-Adaptive Global Registration for Multimodal Retinal Images Using Weakly Supervised Deep-Learning Framework.

Multimodal retinal imaging plays an important role in ophthalmology. We propose a content-adaptive multimodal retinal image registration method in this paper that focuses on the globally coarse alignment and includes three weakly supervised neural networks for vessel segmentation, feature detection and description, and outlier rejection. We apply the proposed framework to register color fundus images with infrared reflectance and fluorescein angiography images, and compare it with several conventional and deep learning methods. Our proposed framework demonstrates a significant improvement in robustness and accuracy reflected by a higher success rate and Dice coefficient compared with other methods.

End-to-End Fovea Localisation in Colour Fundus Images With a Hierarchical Deep Regression Network.

Accurately locating the fovea is a prerequisite for developing computer aided diagnosis (CAD) of retinal diseases. In colour fundus images of the retina, the fovea is a fuzzy region lacking prominent visual features and this makes it difficult to directly locate the fovea. While traditional methods rely on explicitly extracting image features from the surrounding structures such as the optic disc and various vessels to infer the position of the fovea, deep learning based regression technique can implicitly model the relation between the fovea and other nearby anatomical structures to determine the location of the fovea in an end-to-end fashion. Although promising, using deep learning for fovea localisation also has many unsolved challenges. In this paper, we present a new end-to-end fovea localisation method based on a hierarchical coarse-to-fine deep regression neural network. The innovative features of the new method include a multi-scale feature fusion technique and a self-attention technique to exploit location, semantic, and contextual information in an integrated framework, a multi-field-of-view (multi-FOV) feature fusion technique for context-aware feature learning and a Gaussian-shift-cropping method for augmenting effective training data. We present extensive experimental results on two public databases and show that our new method achieved state-of-the-art performances. We also present a comprehensive ablation study and analysis to demonstrate the technical soundness and effectiveness of the overall framework and its various constituent components.

DeepUWF: An Automated Ultra-Wide-Field Fundus Screening System via Deep Learning.

The emerging ultra-wide field of view (UWF) fundus color imaging is a powerful tool for fundus screening. However, manual screening is labor-intensive and subjective. Based on 2644 UWF images, a set of early fundus abnormal screening system named DeepUWF is developed. DeepUWF includes an abnormal fundus screening subsystem and a disease diagnosis subsystem for three kinds of fundus diseases (retinal tear & retinal detachment, diabetic retinopathy and pathological myopia). The components in the system are composed of a set of excellent convolutional neural networks and two custom classifiers. However, the contrast of UWF images used in the research is low, which seriously limits the extraction of fine features of UWF images by depth model. Therefore, the high specificity and low sensitivity of prediction results have always been difficult problems in research. In order to solve this problem, six kinds of image preprocessing techniques are adopted, and their effects on the prediction performance of fundus abnormal and three kinds of fundus diseases models are studied. A variety of experimental indicators are used to evaluate the algorithms for validity and reliability. The experimental results show that these preprocessing methods are helpful to improve the learning ability of the networks and achieve good sensitivity and specificity. Without ophthalmologists, DeepUWF has potential application value, which is helpful for fundus health screening and workflow improvement.

Multimodal imaging of hereditary retinal dystrophies (a series of clinical cases)

Multimodal visualization data of inherited retinal degeneration (IRD) on a Mirante platform (Nidek, Japan), used in a number of clinical cases, is compared with the data obtained by electrophysiological diagnostic methods. 4 patients with varying IRD were examined: adult-onset foveomacular vitelliform dystrophy, Stargardt disease, including those with fundus flavimaculatus, and retinitis pigmentosa. Multimodal imaging includes: colour fundus imaging, fundus autofluorescence, retro mode, and optical coherence tomography. Electroretinography was performed using an MBN electroretinograph (Russia), and electrooculography was performed using a RETIscan Science system (Roland Consult, Germany). Using non-invasive retinal imaging methods, specific patterns of inherited dystrophies were shown, which correlated well with the data of electrophysiological research methods. The combination of multimodal imaging on the Mirante platform (Nidek, Japan) in combination with electrophysiological diagnostic methods can be successfully used in complex diagnostics, monitoring of the progression, and evaluation of the results of IRD treatment.

Identification of Patients with Pentosan Polysulfate Sodium-Associated Maculopathy through Screening of the Electronic Medical Record at an Academic Center.

Aims This chart review of a quaternary academic medical center electronic medical record (EMR) aimed to identify patients at risk of development of maculopathy with exposure to pentosan polysulfate sodium (PPS). Methods A review of electronic medical records of a quaternary medical center of patients with either documented exposure to PPS or diagnosis of interstitial cystitis (IC) from 2007 to 2019 was performed for retinal imaging and visual acuity; the study was conducted in August of 2019. Results 216 charts were included for analysis, of which 96 had documented eye exams and 24 had retinal imaging done. We identified three patients with maculopathy in the context of long-term exposure to PPS via chart review, and one additional patient was identified by referral. The median PPS exposure duration was 11 years (range 7 to 19 years). Median logMAR BCVA OD 0.6 range was 0.0–1.9 (approximate Snellen equivalent 20/80 range (20/20–20/1600)) and OS 0.7 range was 0.1–1.9 (approximate Snellen equivalent 20/100 range (20/25–20/1600)). Ultrawidefield color fundus imaging and fundus autofluorescence revealed findings of pigmentary changes and patchy macular atrophy. Optical coherence tomography (OCT) demonstrated outer retinal thinning and increased choroidal transmission coincident with areas of atrophy seen on fundus imaging. Conclusions Less than half of patients at risk for development of maculopathy due to exposure to PPS had received eye examinations, suggesting that those at risk are not receiving adequate screening. We found two patients with PPS maculopathy who had relatively preserved central vision, one patient with bitemporal vision loss, and one patient who developed vision loss in both eyes.

Adaptive Optics Imaging in Patients Affected by Pseudoxanthoma Elasticum

To describe the retinal findings of patients affected by pseudoxanthoma elasticum (PXE) using a multimodal imaging approach including flood-illumination adaptive optics ophthalmoscopy (AO). Retrospective case series. Patients affected by PXE were retrospectively studied. Clinical data, color, infrared and autofluorescence fundus imaging, optical coherence tomographic scans, and AO examinations were collected. Furthermore, the photoreceptor count was assessed. PXE diagnosis was confirmed by a positive skin biopsy and/or genetic testing. Twenty-one eyes of 18 patients (11 females and 7 males) were included in the study. In 3 patients, both eyes were studied. The mean age at examination was 37 ± 16.4 years (range 11-66) and the mean best-corrected visual acuity (BCVA) was 0.1 ± 0.2 logMAR (range 0-1). We identified 3 types of angioid streaks (AS) using AO: "crack," "band," and "hypopigmented." The first 2 were very similar and they differed in size; the third type showed specific clinical features. Comet lesions appeared as hyper-reflective round lesions on AO imaging. In all eyes, the cone mosaic appeared reduced inside the streaks compared to the neighboring areas (13,532.8 ± 1,366.5 cones/mm2 vs 16,817.1 ± 1,263.0 cones/mm2 respectively). Using AO imaging in PXE-related retinopathy, we were able to observe the presence of the photoreceptors within the angioid streaks, differentiate 3 types of angioid streaks, based on size and reflective features, and identify the very small crystalline bodies not identifiable using other retinal imaging techniques.

Automated diagnosis of diabetic retinopathy enabled by optimized thresholding-based blood vessel segmentation and hybrid classifier

Abstract Objectives The focus of this paper is to introduce an automated early Diabetic Retinopathy (DR) detection scheme from colour fundus images through enhanced segmentation and classification strategies by analyzing blood vessels. Methods The occurrence of DR is increasing from the past years, impacting the eyes due to a sudden rise in the glucose level of blood. All over the world, half of the people who are under age 70 are severely suffered from diabetes. The patients who are affected by DR will lose their vision during the absence of early recognition of DR and appropriate treatment. To decrease the growth and occurrence of loss of vision, the early detection and timely treatment of DR are desirable. At present, deep learning models have presented better performance using retinal images for DR detection. In this work, the input retinal fundus images are initially subjected to pre-processing that undergoes contrast enhancement by Contrast Limited Adaptive Histogram Equalization (CLAHE) and average filtering. Further, the optimized binary thresholding-based segmentation is done for blood vessel segmentation. For the segmented image, Tri-level Discrete Level Decomposition (Tri-DWT) is performed to decompose it. In the feature extraction phase, Local Binary Pattern (LBP), and Gray-Level Co-occurrence Matrices (GLCMs) are extracted. Next, the classification of images is done through the combination of two algorithms, one is Neural Network (NN), and the other Convolutional Neural Network (CNN). The extracted features are subjected to NN, and the tri-DWT-based segmented image is subjected to CNN. Both the segmentation and classification phases are enhanced by the improved meta-heuristic algorithm called Fitness Rate-based Crow Search Algorithm (FR-CSA), in which few parameters are optimized for attaining maximum detection accuracy. Results The proposed DR detection model was implemented in MATLAB 2018a, and the analysis was done using three datasets, HRF, Messidor, and DIARETDB. Conclusions The developed FR-CSA algorithm has the best detection accuracy in diagnosing DR.

Machine Learning Techniques for Ophthalmic Data Processing: A Review.

Machine learning and especially deep learning techniques are dominating medical image and data analysis. This article reviews machine learning approaches proposed for diagnosing ophthalmic diseases during the last four years. Three diseases are addressed in this survey, namely diabetic retinopathy, age-related macular degeneration, and glaucoma. The review covers over 60 publications and 25 public datasets and challenges related to the detection, grading, and lesion segmentation of the three considered diseases. Each section provides a summary of the public datasets and challenges related to each pathology and the current methods that have been applied to the problem. Furthermore, the recent machine learning approaches used for retinal vessels segmentation, and methods of retinal layers and fluid segmentation are reviewed. Two main imaging modalities are considered in this survey, namely color fundus imaging, and optical coherence tomography. Machine learning approaches that use eye measurements and visual field data for glaucoma detection are also included in the survey. Finally, the authors provide their views, expectations and the limitations of the future of these techniques in the clinical practice.

Deep Learning-Based Detection of Early Renal Function Impairment Using Retinal Fundus Images: Model Development and Validation.

BackgroundRetinal imaging has been applied for detecting eye diseases and cardiovascular risks using deep learning–based methods. Furthermore, retinal microvascular and structural changes were found in renal function impairments. However, a deep learning–based method using retinal images for detecting early renal function impairment has not yet been well studied.ObjectiveThis study aimed to develop and evaluate a deep learning model for detecting early renal function impairment using retinal fundus images.MethodsThis retrospective study enrolled patients who underwent renal function tests with color fundus images captured at any time between January 1, 2001, and August 31, 2019. A deep learning model was constructed to detect impaired renal function from the images. Early renal function impairment was defined as estimated glomerular filtration rate <90 mL/min/1.73 m2. Model performance was evaluated with respect to the receiver operating characteristic curve and area under the curve (AUC).ResultsIn total, 25,706 retinal fundus images were obtained from 6212 patients for the study period. The images were divided at an 8:1:1 ratio. The training, validation, and testing data sets respectively contained 20,787, 2189, and 2730 images from 4970, 621, and 621 patients. There were 10,686 and 15,020 images determined to indicate normal and impaired renal function, respectively. The AUC of the model was 0.81 in the overall population. In subgroups stratified by serum hemoglobin A1c (HbA1c) level, the AUCs were 0.81, 0.84, 0.85, and 0.87 for the HbA1c levels of ≤6.5%, >6.5%, >7.5%, and >10%, respectively.ConclusionsThe deep learning model in this study enables the detection of early renal function impairment using retinal fundus images. The model was more accurate for patients with elevated serum HbA1c levels.

Application of an attention U-Net incorporating transfer learning for optic disc and cup segmentation

Optic disc (OD) and optic cup (OC) regional parameters are of utmost importance in the early diagnosis of glaucoma. Improving the accuracy of OD/OC segmentation results and parameter extraction in colour fundus images plays a very important role in early glaucoma screening. To improve the accuracy and inference speed of fundus image segmentation, an algorithm for fundus image segmentation based on an attention U-Net with transfer learning is proposed in this paper. First, an attention gate was added between the encoder and decoder of U-Net to focus on the target areas, thus forming the architecture of the attention U-Net. Then, after the network had been trained on the DRIONS-DB dataset to partially obtain the weights of the encoder, it was trained on the Drishti-GS dataset to further modify the weights. Finally, the trained attention U-Net model incorporating transfer learning was used to segment fundus images. OD/OC extraction using this method shows obvious advantages in model parameter quantity and inference time compared with existing algorithms, the parameter quantity is much smaller than that of existing algorithms, and the model inference time is 0.33 s, representing a reduction of more than 50%. The proposed method can be applied to a fundus image dataset with only a small number of labels. Whilst offering fast OD/OC segmentation, it also guarantees a relatively high segmentation accuracy.

Iterative Augmentation of Visual Evidence for Weakly-Supervised Lesion Localization in Deep Interpretability Frameworks: Application to Color Fundus Images.

Interpretability of deep learning (DL) systems is gaining attention in medical imaging to increase experts' trust in the obtained predictions and facilitate their integration in clinical settings. We propose a deep visualization method to generate interpretability of DL classification tasks in medical imaging by means of visual evidence augmentation. The proposed method iteratively unveils abnormalities based on the prediction of a classifier trained only with image-level labels. For each image, initial visual evidence of the prediction is extracted with a given visual attribution technique. This provides localization of abnormalities that are then removed through selective inpainting. We iteratively apply this procedure until the system considers the image as normal. This yields augmented visual evidence, including less discriminative lesions which were not detected at first but should be considered for final diagnosis. We apply the method to grading of two retinal diseases in color fundus images: diabetic retinopathy (DR) and age-related macular degeneration (AMD). We evaluate the generated visual evidence and the performance of weakly-supervised localization of different types of DR and AMD abnormalities, both qualitatively and quantitatively. We show that the augmented visual evidence of the predictions highlights the biomarkers considered by experts for diagnosis and improves the final localization performance. It results in a relative increase of 11.2± 2.0% per image regarding sensitivity averaged at 10 false positives/image on average, when applied to different classification tasks, visual attribution techniques and network architectures. This makes the proposed method a useful tool for exhaustive visual support of DL classifiers in medical imaging.

Portable ultra-widefield fundus camera for multispectral imaging of the retina and choroid.

Multispectral imaging (MSI) of the retina and choroid has increasing interest for better diagnosis and treatment evaluation of eye diseases. However, currently available MSI systems have a limited field of view (FOV) to evaluate the peripheral retina. This study is to validate trans-pars-planar illumination for a contact-mode ultra-widefield MSI system. By freeing the available pupil for collecting imaging light only, the trans-pars-planar illumination enables a portable, non-mydriatic fundus camera, with 200° FOV in a single fundus image. The trans-pars-planar illumination, delivering illumination light from one side of the eye, naturally enables oblique illumination ophthalmoscopy to enhance the contrast of fundus imaging. A broadband (104 nm) 565 nm light-emitting diode (LED) is used for validating color fundus imaging first. Four narrowband (17-60 nm) 530 nm, 625 nm, 780 nm, and 970 nm LEDs are tested for MSI. With 530 nm illumination, the fundus image reveals retinal vasculature predominantly. 625 nm and 780 nm illuminations enhance the visibility of choroidal vasculature. With further increased wavelength of 970 nm, the fundus image is predominated by large veins in the choroid, with multiple vortex ampullas observed simultaneously in a single fundus image.

Baseline characteristics and age-related macular degeneration in participants of the “ASPirin in Reducing Events in the Elderly” (ASPREE)-AMD trial

PurposeTo describe the baseline participant characteristics in the ASPREE-AMD study, investigating the effect of aspirin on AMD incidence and progression. MethodsAustralian participants from the ASPirin in Reducing Events in the Elderly (ASPREE) trial, randomized to 100 mg aspirin daily or placebo, had non-mydriatic, digital color fundus images graded according to the Beckman AMD classification. Associations with AMD were determined for baseline characteristics and genetic risk variants. ResultsASPREE-AMD sub-study enrolled 4993 participants with gradable macular images. Median age was 73.4 years (IQR, 71.5, 76.6), 52% were female, 10% had diabetes mellitus, 73% had hypertension, and 44% were former/current smokers. Early, intermediate and late AMD (detected in 20.6%, 16.1%, 1.1%, respectively), significantly associated with age, were also associated with increasing HDL levels: OR = 1.52 (95%CI, 1.26, 1.84), OR = 1.43 (1.17, 1.77) and OR = 1.96 (1.02, 3.76), respectively. Female sex was associated with early [OR = 1.37 (1.16, 1.62)], and intermediate [OR = 1.35 (1.12, 1.63)] AMD, as was previous regular use of aspirin, with OR = 1.46 (1.11, 1.92) and OR = 1.37 (1.01, 1.85), respectively. Current smoking had increased odds for late AMD, OR = 4.02 (1.42, 11.36). Genetic risk variant rs3750846 (ARMS2/HTRA1) was associated with each AMD stage (p < 0.001), risk variants rs570618 and rs10922109 (CFH) with intermediate and late AMD (p < 0.001), and rare variant rs147859257 (C3) with late AMD (p < 0.001). The randomized groups were well balanced for all analyzed AMD risk factors. ConclusionsObserved associations are typical of AMD. The ASPREE-AMD clinical trial provides a unique opportunity to determine the risks and benefits of low-dose aspirin for AMD incidence and progression in elderly population. Trial registrationAustralian New Zealand Clinical Trial Registry: ACTRN 12613000755730.

Convexity shape constraints for retinal blood vessel segmentation and foveal avascular zone detection

Diabetic retinopathy (DR) has become a major worldwide health problem due to the increase in blindness among diabetics at early ages. The detection of DR pathologies such as microaneurysms, hemorrhages and exudates through advanced computational techniques is of utmost importance in patient health care. New computer vision techniques are needed to improve upon traditional screening of color fundus images. The segmentation of the entire anatomical structure of the retina is a crucial phase in detecting these pathologies. This work proposes a novel framework for fast and fully automatic blood vessel segmentation and fovea detection. The preprocessing method involved both contrast limited adaptive histogram equalization and the brightness preserving dynamic fuzzy histogram equalization algorithms to enhance image contrast and eliminate noise artifacts. Afterwards, the color spaces and their intrinsic components were examined to identify the most suitable color model to reveal the foreground pixels against the entire background. Several samples were then collected and used by the renowned convexity shape prior segmentation algorithm. The proposed methodology achieved an average vasculature segmentation accuracy exceeding 96%, 95%, 98% and 94% for the DRIVE, STARE, HRF and Messidor publicly available datasets, respectively. An additional validation step reached an average accuracy of 94.30% using an in-house dataset provided by the Hospital Sant Joan of Reus (Spain). Moreover, an outstanding detection accuracy of over 98% was achieved for the foveal avascular zone. An extensive state-of-the-art comparison was also conducted. The proposed approach can thus be integrated into daily clinical practice to assist medical experts in the diagnosis of DR.

Clinical Features of Optic Disc Drusen in an Ophthalmic Genetics Cohort

Materials and Methods Electronic medical records of patients evaluated in the Ophthalmic Genetics clinic at the National Eye Institute (NEI) between 2008 and 2018 were searched for a superficial ODD diagnosis. Color fundus and autofluorescence images were reviewed to confirm ODD, supplemented with optical coherence tomography (OCT) in uncertain cases when available. Demographic information, examination, and genetic testing were reviewed. Disc areas and disc-to-macula distance to disc diameter ratios (DM : DD) were calculated. Results Fifty six of 6207 patients had photographically confirmed ODD (0.9%). Drusen were predominantly bilateral (66%), with a female (62%) and Caucasian (73%) predilection. ODD prevalence in our cohort of patients with inherited retinal degenerations was 2.5%, and ODD were more prevalent in the rod-cone dystrophy subgroup at 2.95% (OR = 3.3 [2.1–5.3], P < 0.001) compared to the ophthalmic genetics cohort. Usher patients were more likely to have ODD (10/132, 7.6%, OR = 9.0 [4.3–17.7], P < 0.001) and had significantly smaller discs compared to the rest of our ODD cohort (disc area: P=0.001, DM : DD: P=0.03). Discussion. While an association between ODD and retinitis pigmentosa has been reported, this study surveys a large cohort of patients with inherited eye conditions and finds the prevalence of superficial ODD is lower than that in the literature. Some subpopulations, such as rod-cone dystrophy and Usher syndrome, had a higher prevalence than the cohort as a whole.

Development and validation of a novel prognostic model for predicting AMD progression using longitudinal fundus images

ObjectiveTo develop a prognostic tool to predict the progression of age-related eye disease progression using longitudinal colour fundus imaging.Methods and analysisPrevious prognostic models using deep learning with imaging data require...

HEREDITARY NEUROPATHIES & ALS: P.107 Review of ophthalmologic findings following intrathecal gene transfer for giant axonal neuropathy

Giant Axonal Neuropathy (GAN) is an ultra-rare childhood onset neurodegenerative disorder of the peripheral and central nervous system. Recessive GAN mutations cause loss of function of gigaxonin, a cytoskeletal regulatory protein, leading to progressive sensorimotor and optic neuropathy, CNS involvement and respiratory failure. We report on the ophthalmologic findings of a single site, phase I, non-randomized, open label dose escalation gene transfer study for GAN (NCT02362438), a first-in-human intrathecal (IT) AAV9 mediated gene transfer. Data from 11 patients with GAN dosed at 3 dose levels; 3.5 × 1013 (1X), 1.2 × 1014 (3.3X), and 1.8 × 1014 (5X) with scAAV9-JeT-GAN with up to 3 years of follow up is analyzed. Ophthalmic evaluations include best-corrected visual acuity, color discrimination thresholds, fundus imaging, and measurement of retinal nerve fiber later (RNFL) thickness through optical coherence tomography. Photopic electroretinography including a photopic negative response and visual evoked potentials (VEP) were also recorded. GAN natural history study data is used for comparison of visual acuity and RNFL thickness as well as correlation of RNFL thickness with validated outcome measures of motor function. Cross sectional natural history analysis shows significant correlation between RNFL thickness and the MFM32 total score. Interim analysis reveals a trend towards stabilization and slowing of the expected decline in RNFL thickness and visual acuity compared to pre-trial lead in data and natural history. VEP results were variable with no consistent trends across natural history and clinical trial cohorts. The interim analysis of ophthalmologic assessments highlights several key aspects: 1) RNFL thickness, a structural assessment of optic nerve atrophy and axonal nerve degeneration at the level of the CNS is a biomarker strongly correlated to overall disease severity. In GAN patients, the decline in RNFL precedes change in visual acuity, and changes can be detected longitudinally. 2) The trend towards stabilization of RNFL thickness post gene transfer suggests that RNFL may be an objective and responsive biomarker of visual system involvement and overall nervous system degeneration in GAN. 3) Assessment of RNFL thickness post gene transfer may be used as an exploratory marker of feasibility of CNS transduction following IT gene transfer.

Glaucoma assessment from color fundus images using convolutional neural network

AbstractEarly detection and proper screening are essential to prevent vision loss due to glaucoma. In recent years, convolutional neural network (CNN) has been successfully applied to the color fundus images for the automatic detection of glaucoma. Compared to the existing automatic screening methods, CNNs have the ability to extract the distinct features directly from the fundus images. In this paper, a deep learning architecture based on a CNN is designed for the classification of glaucomatous and normal fundus images. An 18–layer CNN is designed and trained to extract the discriminative features from the fundus image. It comprises of four convolutional layers, two max pooling layers, and one fully connected layer. A two–stage tuning approach is proposed for the selection of suitable batch size and initial learning rate. The proposed network is tested on DRISHTI–GS1, ORIGA, RIM–ONE2 (release 2), ACRIMA, and large–scale attention–based glaucoma (LAG) databases. Rotation data augmentation technique is employed to enlarge the dataset. Randomly selected 70% of images are used for training the model and remaining 30% images are used for testing. An overall accuracy of 86.62%, 85.97%, 78.32%, 94.43%, and 96.64% are obtained on DRISHTI–GS1, RIM–ONE2, ORIGA, LAG, and ACRIMA databases, respectively. The proposed method has achieved an accuracy, sensitivity, specificity, and precision of 96.64%, 96.07%, 97.39%, and 97.74%, respectively, for ACRIMA database. Compared to other existing architectures, the proposed method is robust to Gaussian noise and salt–and–pepper noise.