Corneal guttae, which are the abnormal growth of extracellular matrix in the corneal endothelium, are observed in specular images as black droplets that occlude the endothelial cells. To estimate the corneal parameters (endothelial cell density [ECD], coefficient of variation [CV], and hexagonality [HEX]), we propose a new deep learning method that includes a novel attention mechanism (named fNLA), which helps to infer the cell edges in the occluded areas. The approach first derives the cell edges, then infers the well-detected cells, and finally employs a postprocessing method to fix mistakes. This results in a binary segmentation from which the corneal parameters are estimated. We analyzed 1203 images (500 contained guttae) obtained with a Topcon SP-1P microscope. To generate the ground truth, we performed manual segmentation in all images. Several networks were evaluated (UNet, ResUNeXt, DenseUNets, UNet++, etc.) and we found that DenseUNets with fNLA provided the lowest error: a mean absolute error of 23.16 [cells/mm²] in ECD, 1.28 [%] in CV, and 3.13 [%] in HEX. Compared with Topcon’s built-in software, our error was 3–6 times smaller. Overall, our approach handled notably well the cells affected by guttae, detecting cell edges partially occluded by small guttae and discarding large areas covered by extensive guttae.

@en

Purpose: To present a fully automatic method to estimate the corneal endothelium parameters from specular microscopy images and to use it to study a one-year follow-up after ultrathin Descemet stripping automated endothelial keratoplasty. Methods: We analyzed 383 post ultrathin Descemet stripping automated endothelial keratoplasty images from 41 eyes acquired with a Topcon SP-1P specular microscope at 1, 3, 6, and 12 months after surgery. The estimated parameters were endothelial cell density (ECD), coefficient of variation (CV), and hexagonality (HEX). Manual segmentation was performed in all images. Results: Our method provided an estimate for ECD, CV, and HEX in 98.4% of the images, whereas Topcon’s software had a success rate of 71.5% for ECD/CV and 30.5% for HEX. For the images with estimates, the percentage error in our method was 2.5% for ECD, 5.7% for CV, and 5.7% for HEX, whereas Topcon’s software provided an error of 7.5% for ECD, 17.5% for CV, and 18.3% for HEX. Our method was significantly better than Topcon’s (P < 0.0001) and was not statistically significantly different from the manual assessments (P > 0.05). At month 12, the subjects presented an average ECD = 1377 ± 483 [cells/mm² ], CV = 26.1 ± 5.7 [%], and HEX = 58.1 ± 7.1 [%]. Conclusions: The proposed method obtains reliable and accurate estimations even in challenging specular images of pathologic corneas. Translational Relevance: CV and HEX, not currently used in the clinic owing to a lack of reliability in automatic methods, are useful biomarkers to analyze the postoperative healing process. Our accurate estimations allow now for their clinical use.

@en

PURPOSE: To study the disease course of RPE65-associated inherited retinal degenerations (IRDs) as a function of the genotype, define a critical age for blindness, and identify potential modifiers. METHODS: Forty-five patients with IRD from 33 families with biallelic RPE65 mutations, 28 stemming from a genetic isolate. We collected retrospective data from medical charts. Coexisting variants in 108 IRD-associated genes were identified with Molecular Inversion Probe analysis. RESULTS: Most patients were diagnosed within the first years of life. Daytime visual function ranged from near-normal to blindness in the first four decades and met WHO criteria for blindness for visual acuity and visual field in the fifth decade. p.(Thr368His) was the most common variant (54%). Intrafamilial variability and interfamilial variability in disease severity and progression were observed. Molecular Inversion Probe analysis confirmed all RPE65 variants and identified one additional variant in LRAT and one in EYS in two separate patients. CONCLUSION: All patients with RPE65-associated IRDs developed symptoms within the first year of life. Visual function in childhood and adolescence varied but deteriorated inevitably toward blindness after age 40. In this study, genotype was not predictive of clinical course. The variance in severity of disease could not be explained by double hits in other IRD genes.

@en

In images of the corneal endothelium (CE) acquired by specular microscopy, endothelial cells are commonly only visible in a part of the image due to varying contrast, mainly caused by challenging imaging conditions as a result of a strongly curved endothelium. In order to estimate the morphometric parameters of the corneal endothelium, the analyses need to be restricted to trustworthy regions - the region of interest (ROI) - where individual cells are discernible. We developed an automatic method to find the ROI by Dense U-nets, a densely connected network of convolutional layers. We tested the method on a heterogeneous dataset of 140 images, which contains a large number of blurred, noisy, and/or out of focus images, where the selection of the ROI for automatic biomarker extraction is vital. By using edge images as input, which can be estimated after retraining the same network, Dense U-net detected the trustworthy areas with an accuracy of 98.94% and an area under the ROC curve (AUC) of 0.998, without being affected by the class imbalance (9:1 in our dataset). After applying the estimated ROI to the edge images, the mean absolute percentage error (MAPE) in the estimated endothelial parameters was 0.80% for ECD, 3.60% for CV, and 2.55% for HEX.

@en

The morphometric parameters of the corneal endothelium – cell density (ECD), cell size variation (CV), and hexagonality (HEX) – provide clinically relevant information about the cornea. To estimate these parameters, the endothelium is commonly imaged with a non-contact specular microscope and cell segmentation is performed to these images. In previous work, we have developed several methods that, combined, can perform an automated estimation of the parameters: the inference of the cell edges, the detection of the region of interest (ROI), a post-processing method that combines both images (edges and ROI), and a refinement method that removes false edges. In this work, we first explore the possibility of using a CNN-based regressor to directly infer the parameters from the edge images, simplifying the framework. We use a dataset of 738 images coming from a study related to the implantation of a Baerveldt glaucoma device and a standard clinical care regarding DSAEK corneal transplantation, both from the Rotterdam Eye Hospital and both containing images of unhealthy endotheliums. This large dataset allows us to build a large training set that makes this approach feasible. We achieved a mean absolute percentage error (MAPE) of 4.32% for ECD, 7.07% for CV, and 11.74% for HEX. These results, while promising, do not outperform our previous work. In a second experiment, we explore the use of the CNN-based regressor to improve the post-processing method of our previous approach in order to adapt it to the specifics of each image. Our results showed no clear benefit and proved that our previous post-processing is already highly reliable and robust.@en

Background

Corneal endothelium (CE) images provide valuable clinical information regarding the health state of the cornea. Computation of the clinical morphometric parameters requires the segmentation of endothelial cell images. Current techniques to image the endothelium in vivo deliver low quality images, which makes automatic segmentation a complicated task. Here, we present two convolutional neural networks (CNN) to segment CE images: a global fully convolutional approach based on U-net, and a local sliding-window network (SW-net). We propose to use probabilistic labels instead of binary, we evaluate a preprocessing method to enhance the contrast of images, and we introduce a postprocessing method based on Fourier analysis and watershed to convert the CNN output images into the final cell segmentation. Both methods are applied to 50 images acquired with an SP-1P Topcon specular microscope. Estimates are compared against a manual delineation made by a trained observer.

Results

U-net (AUC=0.9938) yields slightly sharper, clearer images than SW-net (AUC=0.9921). After postprocessing, U-net obtains a DICE=0.981 and a MHD=0.22 (modified Hausdorff distance), whereas SW-net yields a DICE=0.978 and a MHD=0.30. U-net generates a wrong cell segmentation in only 0.48% of the cells, versus 0.92% for the SW-net. U-net achieves statistically significant better precision and accuracy than both, Topcon and SW-net, for the estimates of three clinical parameters: cell density (ECD), polymegethism (CV), and pleomorphism (HEX). The mean relative error in U-net for the parameters is 0.4% in ECD, 2.8% in CV, and 1.3% in HEX. The computation time to segment an image and estimate the parameters is barely a few seconds.

Conclusions

Both methods presented here provide a statistically significant improvement over the state of the art. U-net has reached the smallest error rate. We suggest a segmentation refinement based on our previous work to further improve the performance.@en

The attenuation coefficient (AC) is a property related to the microstructure of tissue on a wavelength scale that can be estimated from optical coherence tomography (OCT) data. Since the OCT signal sensitivity is affected by the finite spectrometer/detector resolution called roll-off and the shape of the focused beam in the sample arm, ignoring these effects leads to severely biased estimates of AC. Previously, the signal intensity dependence on these factors has been modeled. In this paper, we study the dependence of the estimated AC on the beam-shape and focus depth experimentally. A method is presented to estimate the axial point spread function model parameters by fitting the OCT signal model for single scattered light to the averaged A-lines of multiple B-scans obtained from a homogeneous single-layer phantom. The estimated model parameters were used to compensate the signal for the axial point spread function and roll-off in order to obtain an accurate estimate of AC. The result shows a significant improvement in the accuracy of the estimation of AC after correcting for the shape of the OCT beam.@en

Thinning of the retinal nerve fiber layer (RNFL) is a recently discovered feature of Parkinson’s disease (PD). Its exact pathological mechanism is yet unknown. We aimed to determine whether morphological changes of the RNFL are limited to RNFL thinning or also comprise an altered internal structure of this layer. Therefore, we investigated RNFL thickness and applied the RNFL attenuation coefficient (RNFL-AC), a novel method derived from optical coherence tomography, in PD patients and healthy controls (HCs). In this pilot study, we included 20 PD patients and 20 HCs matched for age, sex, and ethnicity. An ophthalmologist investigated all participants thoroughly, and we acquired retinal images from both eyes of each participant with a Spectralis optical coherence tomography system. We obtained both the RNFL-AC and RNFL thickness from peripapillary RNFL scans for the entire RNFL, as well as for each quadrant separately. We found no significant differences in the average RNFL-AC or the RNFL-AC of the separate retinal quadrants between PD patients and the HC group. However, compared to the HC group, PD patients had a significantly thinner RNFL in the temporal retinal quadrant. RNFL thinning was found in the temporal quadrant in PD patients without a corresponding change in the RNFL-AC. These findings suggest a reduction in the number of RNFL axons (atrophy) without other major changes in the structural integrity of the remaining RNFL.

@en

Clinical parameters related to the corneal endothelium can only be estimated by segmenting endothelial cell images. Specular microscopy is the current standard technique to image the endothelium, but its low SNR make the segmentation a complicated task. Recently, we proposed a method to segment such images by starting with an oversegmented image and merging the superpixels that constitute a cell. Here, we show how our merging method provides better results than optimizing the segmentation itself. Furthermore, our method can provide accurate results despite the degree of the initial oversegmentation, resulting into a precision and recall of 0.91 for the optimal oversegmentation.

@en

Corneal endothelium images obtained by in vivo specular microscopy provide important information to assess the health status of the cornea. Estimation of clinical parameters, such as cell density, polymegethism, and pleomorphism, requires accurate cell segmentation. State-of-the-art techniques to automatically segment the endothelium are error-prone when applied to images with low contrast and/or large variation in cell size. Here, we propose an automatic method to segment the endothelium. Starting with an oversegmented image comprised of superpixels obtained from a stochastic watershed segmentation, the proposed method uses intensity and shape information of the superpixels to identify and merge those that constitute a cell, using Support Vector Machines. We evaluated the automatic segmentation on a dataset of in vivo specular microscopy images (Topcon SP-1P), obtaining 95.8merged cells and 2.0the parameter estimation against the results of the vendor&#x2019;s builtin software, obtaining a statistically significant better precision in all parameters and a similar or better accuracy. The parameter estimation was also evaluated on three other datasets from different imaging modalities (confocal microscopy, phasecontrast microscopy, and fluorescence confocal microscopy) and tissue types (ex vivo corneal endothelium and retinal pigment epithelium). In comparison with the estimates of the datasets&#x2019; authors, we achieved statistically significant better accuracy and precision in all parameters except pleomorphism, where a similar accuracy and precision were obtained.

@en

Automated detection and quantification of spatio-temporal retinal changes is an important step to objectively assess disease progression and treatment effects for dynamic retinal diseases such as diabetic retinopathy (DR). However, detecting retinal changes caused by early DR lesions such as microaneurysms and dot hemorrhages from longitudinal pairs of fundus images is challenging due to intra and inter-image illumination variation between fundus images. This paper explores a method for automated detection of retinal changes from illumination normalized fundus images using a deep convolutional neural network (CNN), and compares its performance with two other CNNs trained separately on color and green channel fundus images. Illumination variation was addressed by correcting for the variability in the luminosity and contrast estimated from a large scale retinal regions. The CNN models were trained and evaluated on image patches extracted from a registered fundus image set collected from 51 diabetic eyes that were screened at two different time-points. The results show that using normalized images yield better performance than color and green channel images, suggesting that illumination normalization greatly facilitates CNNs to quickly and correctly learn distinctive local image features of DR related retinal changes.

@en

People with diabetes mellitus need annual screening to check for the development of diabetic retinopathy. Tracking small retinal changes due to early diabetic retinopathy lesions in longitudinal fundus image sets is challenging due to intra- and inter-visit variability in illumination and image quality, the required high registration accuracy, and the subtle appearance of retinal lesions compared to other retinal features. This paper presents a robust and flexible approach for automated detection of longitudinal retinal changes due to small red lesions by exploiting normalized fundus images that significantly reduce illumination variations and improve the contrast of small retinal features. To detect spatio-temporal retinal changes, the absolute difference between the extremes of the multiscale blobness responses of fundus images from two time-points is proposed as a simple and effective blobness measure. DR related changes are then identified based on several intensity and shape features by a support vector machine classifier. The proposed approach was evaluated in the context of a regular diabetic retinopathy screening program involving subjects ranging from healthy (no retinal lesion) to moderate (with clinically relevant retinal lesions) DR levels. Evaluation shows that the system is able to detect retinal changes due to small red lesions with a sensitivity of 80% at an average false positive rate of 1 and 2.5 lesions per eye on small and large fields-of-view of the retina, respectively.

@en

Optical coherence tomography (OCT) yields high-resolution, three-dimensional images of the retina. A better understanding of retinal nerve fiber bundle (RNFB) trajectories in combination with visual field data may be used for future diagnosis and monitoring of glaucoma. However, manual tracing of these bundles is a tedious task. In this work, we present an automatic technique to estimate the orientation of RNFBs from volumetric OCT scans. Our method consists of several steps, starting from automatic segmentation of the RNFL. Then, a stack of en face images around the posterior nerve fiber layer interface was extracted. The image showing the best visibility of RNFB trajectories was selected for further processing. After denoising the selected en face image, a semblance structure-oriented filter was applied to probe the strength of local linear structure in a discrete set of orientations creating an orientation space. Gaussian filtering along the orientation axis in this space is used to find the dominant orientation. Next, a confidence map was created to supplement the estimated orientation. This confidence map was used as pixel weight in normalized convolution to regularize the semblance filter response after which a new orientation estimate can be obtained. Finally, after several iterations an orientation field corresponding to the strongest local orientation was obtained. The RNFB orientations of six macular scans from three subjects were estimated. For all scans, visual inspection shows a good agreement between the estimated orientation fields and the RNFB trajectories in the en face images. Additionally, a good correlation between the orientation fields of two scans of the same subject was observed. Our method was also applied to a larger field of view around the macula. Manual tracing of the RNFB trajectories shows a good agreement with the automatically obtained streamlines obtained by fiber tracking.

@en

We present a locally-adaptive approach to segment the fluid and the interfaces between retinal layers in eyes affected by central serous retinopathy based on loosely-coupled level sets. The approach exploits the local attenuation coefficient differences of layers around an interface and introduces auxiliary interfaces to delineate the fluid. Thus, it can handle abrupt attenuation coefficient variations and topology-disrupting anomalies. The method was applied to in-vivo images of retinas acquired by optical coherence tomography. A quantitative comparison with manual annotations shows the method's high accuracy: we obtained a mean absolute deviation for the interfaces of 3.7-8.9 βm (1-2 pixels) and a Dice coefficient for the fluid segmentation of 0.96.

@en

Optical coherence tomography (OCT) is used to produce high-resolution three-dimensional images of the retina, which permit the investigation of retinal irregularities. In dry age-related macular degeneration (AMD), a chronic eye disease that causes central vision loss, disruptions such as drusen and changes in retinal layer thicknesses occur which could be used as biomarkers for disease monitoring and diagnosis. Due to the topology disrupting pathology, existing segmentation methods often fail. Here, we present a solution for the segmentation of retinal layers in dry AMD subjects by extending our previously presented loosely coupled level sets framework which operates on attenuation coefficients. In eyes affected by AMD, Bruch’s membrane becomes visible only below the drusen and our segmentation framework is adapted to delineate such a partially discernible interface. Furthermore, the initialization stage, which tentatively segments five interfaces, is modified to accommodate the appearance of drusen. This stage is based on Dijkstra's algorithm and combines prior knowledge on the shape of the interface, gradient and attenuation coefficient in the newly proposed cost function. This prior knowledge is incorporated by varying the weights for horizontal, diagonal and vertical edges. Finally, quantitative evaluation of the accuracy shows a good agreement between manual and automated segmentation. @en