Retinal layer segmentation of macular OCT images using boundary classification

Retinal OCT image classification based on MGR-GAN

Accurately classifying optical coherence tomography (OCT) images is essential for diagnosing and treating ophthalmic diseases. This paper introduces a novel generative adversarial network framework called MGR-GAN. The masked image modeling (MIM) method is integrated into the GAN model's generator, enhancing its ability to synthesize more realistic images by reconstructing them based on unmasked patches. A ResNet-structured discriminator is employed to determine whether the image is generated by the generator. Through the unique game process of the generative adversarial network (GAN) model, the discriminator acquires high-level discriminant features, essential for precise OCT classification. Experimental results demonstrate that MGR-GAN achieves a classification accuracy of 98.4% on the original UCSD dataset. As the trained generator can synthesize OCT images with higher precision, and owing to category imbalances in the UCSD dataset, the generated OCT images are leveraged to address this imbalance. After balancing the UCSD dataset, the classification accuracy further improves to 99%.

A weak edge estimation based multi-task neural network for OCT segmentation

Optical Coherence Tomography (OCT) offers high-resolution images of the eye's fundus. This enables thorough analysis of retinal health by doctors, providing a solid basis for diagnosis and treatment. With the development of deep learning, deep learning-based methods are becoming more popular for fundus OCT image segmentation. Yet, these methods still encounter two primary challenges. Firstly, deep learning methods are sensitive to weak edges. Secondly, the high cost of annotating medical image data results in a lack of labeled data, leading to overfitting during model training. To tackle these challenges, we introduce the Multi-Task Attention Mechanism Network with Pruning (MTAMNP), consisting of a segmentation branch and a boundary regression branch. The boundary regression branch utilizes an adaptive weighted loss function derived from the Truncated Signed Distance Function(TSDF), improving the model's capacity to preserve weak edge details. The Spatial Attention Based Dual-Branch Information Fusion Block links these branches, enabling mutual benefit. Furthermore, we present a structured pruning method grounded in channel attention to decrease parameter count, mitigate overfitting, and uphold segmentation accuracy. Our method surpasses other cutting-edge segmentation networks on two widely accessible datasets, achieving Dice scores of 84.09% and 93.84% on the HCMS and Duke datasets.

Boundary-Repairing Dual-Path Network for Retinal Layer Segmentation in OCT Image with Pigment Epithelial Detachment

Automatic retinal layer segmentation in optical coherence tomography (OCT) images is crucial for the diagnosis of ocular diseases. Currently, automatic retinal layer segmentation works well with normal OCT images. However, pigment epithelial detachment (PED) dramatically alters the retinal structure, causing blurred boundaries and partial disappearance of the Bruch's Membrane (BM), thus posing challenges to the segmentation. To tackle these problems, we propose a novel dual-path U-shaped network for simultaneous layer segmentation and boundary regression. This network first designs a feature interaction fusion (FIF) module to strengthen the boundary shape constraints in the layer path. To address the challenge posed by partial BM disappearance and boundary-blurring, we propose a layer boundary repair (LBR) module. This module aims to use contrastive loss to enhance the confidence of blurred boundary regions and refine the segmentation of layer boundaries through the re-prediction head. In addition, we introduce a novel bilateral threshold distance map (BTDM) designed for the boundary path. The BTDM serves to emphasize information within boundary regions. This map, combined with the updated probability map, culminates in topology-guaranteed segmentation results achieved through a topology correction (TC) module. We investigated the proposed network on two severely deformed datasets (i.e., OCTA-500 and Aier-PED) and one slightly deformed dataset (i.e., DUKE). The proposed method achieves an average Dice score of 94.26% on the OCTA-500 dataset, which was 1.5% higher than BAU-Net and outperformed other methods. In the DUKE and Aier-PED datasets, the proposed method achieved average Dice scores of 91.65% and 95.75%, respectively.

Inner retinal layer thickness alterations in adult and pediatric patients with neurofibromatosis 1

This study compared the thickness of each intraretinal layer in patients with neurofibromatosis 1 (NF1) and controls to analyze the association between intraretinal layer thickness and visual function. The macular spectral-domain optical coherence tomography volumetric dataset obtained from 68 eyes (25 adult eyes, 43 pediatric eyes) with NF1 without optic glioma and 143 control eyes (100 adult eyes, 43 pediatric eyes) was used for image auto-segmentation. The intraretinal layers segmented from the volumetric data included the macular retinal nerve fiber layer (RNFL), ganglion cell-inner plexiform layer (GCIPL), inner nuclear layer, outer plexiform layer, outer nuclear layer, and photoreceptor layer. Cases and controls were compared after adjusting for age, sex, refractive error, and binocular use. The association between retinal layer thickness and visual acuity was also analyzed. The GCIPL was significantly thinner in both adult and pediatric patients with NF1 compared with healthy controls. Average RNFL and GCIPL thicknesses were associated with visual acuity in adult patients with NF1. In pediatric patients, average GCIPL thickness was associated with visual acuity. These results suggest that changes in the inner retinal layer could be a biomarker of the structural and functional status of patients with NF1.

Automatic segmentation of layers in chorio-retinal complex using Graph-based method for ultra-speed 1.7 MHz wide field swept source FDML optical coherence tomography

The posterior segment of the human eye complex contains two discrete microstructure and vasculature network systems, namely, the retina and choroid. We present a single segmentation framework technique for segmenting the entire layers present in the chorio-retinal complex of the human eye using optical coherence tomography (OCT) images. This automatic program is based on the graph theory method. This single program is capable of segmenting seven layers of the retina and choroid scleral interface. The graph theory was utilized to find the probability matrix and subsequent boundaries of different layers. The program was also implemented to segment angiographic maps of different chorio-retinal layers using "segmentation matrices." The method was tested and successfully validated on OCT images from six normal human eyes as well as eyes with non-exudative age-related macular degeneration (AMD). The thickness of microstructure and microvasculature for different layers located in the chorio-retinal segment of the eye was also generated and compared. A decent efficiency in terms of processing time, sensitivity, and accuracy was observed compared to the manual segmentation and other existing methods. The proposed method automatically segments whole OCT images of chorio-retinal complex with augmented probability maps generation in OCT volume dataset. We have also evaluated the segmentation results using quantitative metrics such as Dice coefficient and Hausdorff distance This method realizes a mean descent Dice similarity coefficient (DSC) value of 0.82 (range, 0.816-0.864) for RPE and CSI layer.

Advancing Ocular Imaging: A Hybrid Attention Mechanism-Based U-Net Model for Precise Segmentation of Sub-Retinal Layers in OCT Images

This paper presents a novel U-Net model incorporating a hybrid attention mechanism for automating the segmentation of sub-retinal layers in Optical Coherence Tomography (OCT) images. OCT is an ophthalmology tool that provides detailed insights into retinal structures. Manual segmentation of these layers is time-consuming and subjective, calling for automated solutions. Our proposed model combines edge and spatial attention mechanisms with the U-Net architecture to improve segmentation accuracy. By leveraging attention mechanisms, the U-Net focuses selectively on image features. Extensive evaluations using datasets demonstrate that our model outperforms existing approaches, making it a valuable tool for medical professionals. The study also highlights the model's robustness through performance metrics such as an average Dice score of 94.99%, Adjusted Rand Index (ARI) of 97.00%, and Strength of Agreement (SOA) classifications like "Almost Perfect", "Excellent", and "Very Strong". This advanced predictive model shows promise in expediting processes and enhancing the precision of ocular imaging in real-world applications.

Retinal Layer Segmentation in OCT Images With Boundary Regression and Feature Polarization

The geometry of retinal layers is an important imaging feature for the diagnosis of some ophthalmic diseases. In recent years, retinal layer segmentation methods for optical coherence tomography (OCT) images have emerged one after another, and huge progress has been achieved. However, challenges due to interference factors such as noise, blurring, fundus effusion, and tissue artifacts remain in existing methods, primarily manifesting as intra-layer false positives and inter-layer boundary deviation. To solve these problems, we propose a method called Tightly combined Cross-Convolution and Transformer with Boundary regression and feature Polarization (TCCT-BP). This method uses a hybrid architecture of CNN and lightweight Transformer to improve the perception of retinal layers. In addition, a feature grouping and sampling method and the corresponding polarization loss function are designed to maximize the differentiation of the feature vectors of different retinal layers, and a boundary regression loss function is devised to constrain the retinal boundary distribution for a better fit to the ground truth. Extensive experiments on four benchmark datasets demonstrate that the proposed method achieves state-of-the-art performance in dealing with problems of false positives and boundary distortion. The proposed method ranked first in the OCT Layer Segmentation task of GOALS challenge held by MICCAI 2022. The source code is available at https://www.github.com/tyb311/TCCT.

Rapid measurement of epidermal thickness in OCT images of skin

Epidermal thickness (ET) changes are associated with several skin diseases. To measure ET, segmentation of optical coherence tomography (OCT) images is essential; manual segmentation is very time-consuming and requires training and some understanding of how to interpret OCT images. Fast results are important in order to analyze ET over different regions of skin in rapid succession to complete a clinical examination and enable the physician to discuss results with the patient in real time. The well-known CNN-graph search (CNN-GS) methodology delivers highly accurate results, but at a high computational cost. Our objective was to build a computational core, based on CNN-GS, able to accurately segment OCT skin images in real time. We accomplished this by fine-tuning the hyperparameters, testing a range of speed-up algorithms including pruning and quantization, designing a novel pixel-skipping process, and implementing the final product with efficient use of core and threads on a multicore central processing unit (CPU). We name this product CNN-GS-skin. The method identifies two defined boundaries on OCT skin images in order to measure ET. We applied CNN-GS-skin to OCT skin images, taken from various body sites of 63 healthy individuals. Compared with CNN-GS, our described method reduced computation time by 130 [Formula: see text] with minimal reduction in ET determination accuracy (from 96.38 to 94.67%).

Simultaneous alignment and surface regression using hybrid 2D-3D networks for 3D coherent layer segmentation of retinal OCT images with full and sparse annotations

Layer segmentation is important to quantitative analysis of retinal optical coherence tomography (OCT). Recently, deep learning based methods have been developed to automate this task and yield remarkable performance. However, due to the large spatial gap and potential mismatch between the B-scans of an OCT volume, all of them were based on 2D segmentation of individual B-scans, which may lose the continuity and diagnostic information of the retinal layers in 3D space. Besides, most of these methods required dense annotation of the OCT volumes, which is labor-intensive and expertise-demanding. This work presents a novel framework based on hybrid 2D-3D convolutional neural networks (CNNs) to obtain continuous 3D retinal layer surfaces from OCT volumes, which works well with both full and sparse annotations. The 2D features of individual B-scans are extracted by an encoder consisting of 2D convolutions. These 2D features are then used to produce the alignment displacement vectors and layer segmentation by two 3D decoders coupled via a spatial transformer module. Two losses are proposed to utilize the retinal layers' natural property of being smooth for B-scan alignment and layer segmentation, respectively, and are the key to the semi-supervised learning with sparse annotation. The entire framework is trained end-to-end. To the best of our knowledge, this is the first work that attempts 3D retinal layer segmentation in volumetric OCT images based on CNNs. Experiments on a synthetic dataset and three public clinical datasets show that our framework can effectively align the B-scans for potential motion correction, and achieves superior performance to state-of-the-art 2D deep learning methods in terms of both layer segmentation accuracy and cross-B-scan 3D continuity in both fully and semi-supervised settings, thus offering more clinical values than previous works.

Graph Attention U-Net for Retinal Layer Surface Detection and Choroid Neovascularization Segmentation in OCT Images

Choroidal neovascularization (CNV) is a typical symptom of age-related macular degeneration (AMD) and is one of the leading causes for blindness. Accurate segmentation of CNV and detection of retinal layers are critical for eye disease diagnosis and monitoring. In this paper, we propose a novel graph attention U-Net (GA-UNet) for retinal layer surface detection and CNV segmentation in optical coherence tomography (OCT) images. Due to retinal layer deformation caused by CNV, it is challenging for existing models to segment CNV and detect retinal layer surfaces with the correct topological order. We propose two novel modules to address the challenge. The first module is a graph attention encoder (GAE) in a U-Net model that automatically integrates topological and pathological knowledge of retinal layers into the U-Net structure to achieve effective feature embedding. The second module is a graph decorrelation module (GDM) that takes reconstructed features by the decoder of the U-Net as inputs, it then decorrelates and removes information unrelated to retinal layer for improved retinal layer surface detection. In addition, we propose a new loss function to maintain the correct topological order of retinal layers and the continuity of their boundaries. The proposed model learns graph attention maps automatically during training and performs retinal layer surface detection and CNV segmentation simultaneously with the attention maps during inference. We evaluated the proposed model on our private AMD dataset and another public dataset. Experiment results show that the proposed model outperformed the competing methods for retinal layer surface detection and CNV segmentation and achieved new state of the arts on the datasets.

Automatic Segmentation of Retinal Layers in Multiple Neurodegenerative Disorder Scenarios

Retinal Optical Coherence Tomography (OCT) allows the non-invasive direct observation of the central nervous system, enabling the measurement and extraction of biomarkers from neural tissue that can be helpful in the assessment of ocular, systemic and Neurological Disorders (ND). Deep learning models can be trained to segment the retinal layers for biomarker extraction. However, the onset of ND can have an impact on the neural tissue, which can lead to the degraded performance of models not exposed to images displaying signs of disease during training. We present a fully automatic approach for the retinal layer segmentation in multiple neurodegenerative disorder scenarios, using an annotated dataset of patients of the most prevalent NDs: Alzheimer's disease, Parkinson's disease, multiple sclerosis and essential tremor, along with healthy control patients. Furthermore, we present a two-part, comprehensive study on the effects of ND on the performance of these models. The results show that images of healthy patients may not be sufficient for the robust training of automated segmentation models intended for the analysis of ND patients, and that using images representative of different NDs can increase the model performance. These results indicate that the presence or absence of patients of ND in datasets should be taken into account when training deep learning models for retinal layer segmentation, and that the proposed approach can provide a valuable tool for the robust and reliable diagnosis in multiple scenarios of ND.

Recurrent Self Fusion: Iterative Denoising for Consistent Retinal OCT Segmentation

Optical coherence tomography (OCT) is a valuable imaging technique in ophthalmology, providing high-resolution, cross-sectional images of the retina for early detection and monitoring of various retinal and neurological diseases. However, discrepancies in retinal layer thickness measurements among different OCT devices pose challenges for data comparison and interpretation, particularly in longitudinal analyses. This work introduces the idea of a recurrent self fusion (RSF) algorithm to address this issue. Our RSF algorithm, built upon the self fusion methodology, iteratively denoises retinal OCT images. A deep learning-based retinal OCT segmentation algorithm is employed for downstream analyses. A large dataset of paired OCT scans acquired on both a Spectralis and Cirrus OCT device are used for validation. The results demonstrate that the RSF algorithm effectively reduces speckle contrast and enhances the consistency of retinal OCT segmentation.

Effects of Ibudilast on Retinal Atrophy in Progressive Multiple Sclerosis Subtypes: Post Hoc Analyses of the SPRINT-MS Trial

BACKGROUND AND OBJECTIVES

Ganglion cell + inner plexiform layer (GCIPL) thinning, measured by optical coherence tomography (OCT), reflects global neurodegeneration in multiple sclerosis (MS). Atrophy of the inner (INL) and outer nuclear layer (ONL) may also be prominent in progressive MS (PMS). The phase 2, SPRINT-MS trial found reduced brain atrophy with ibudilast therapy in PMS. In this post hoc analysis of the SPRINT-MS trial, we investigate (1) retinal atrophy (2) differences in response by subtype and (3) associations between OCT and MRI measures of neurodegeneration.

METHODS

In the multicenter, double-blind SPRINT-MS trial, participants with secondary progressive MS (SPMS) or primary progressive MS (PPMS) were randomized to ibudilast or placebo. OCT and MRI data were collected every 24 weeks for 96 weeks. Extensive OCT quality control and algorithmic segmentation produced consistent results across Cirrus HD-OCT and Spectralis devices. Primary endpoints were GCIPL, INL, and ONL atrophy, assessed by linear mixed-effects regression. Secondary endpoints were associations of OCT measures, brain parenchymal fraction, and cortical thickness, assessed by partial Pearson correlations.

RESULTS

One hundred thirty-four PPMS and 121 SPMS participants were included. GCIPL atrophy was 79% slower in the ibudilast (-0.07 ± 0.23 µm/y) vs placebo group (-0.32 ± 0.20 µm/y, p = 0.003). This effect predominated in the PPMS cohort (ibudilast: -0.08 ± 0.29 µm/y vs placebo: -0.60 ± 0.29 µm/y, a decrease of 87%, p < 0.001) and was not detected in the SPMS cohort (ibudilast: -0.21 ± 0.28 µm/y vs placebo: -0.14 ± 0.27 µm/y, p = 0.55). GCIPL, INL, and ONL atrophy rates correlated with whole brain atrophy rates across the cohort (r = 0.27, r = 0.26, and r = 0.20, respectively; p < 0.001). Power calculations from these data show future trials of similar size and design have ≥80% power to detect GCIPL atrophy effect sizes of approximately 40%.

DISCUSSION

Ibudilast treatment decreased GCIPL atrophy in PMS, driven by the PPMS cohort, with no effect seen in SPMS. Modulated atrophy of retinal layers may be detectable in sample sizes smaller than the SPRINT-MS trial and correlate with whole brain atrophy in PMS, further highlighting their utility as outcomes in PMS.

CLASSIFICATION OF EVIDENCE

This study provides Class II evidence that ibudilast reduces composite ganglion cell + inner plexiform layer atrophy, without reduction of inner or outer nuclear layer atrophy, in patients with primary progressive MS but not those with secondary progressive MS.

Deep learning network with differentiable dynamic programming for retina OCT surface segmentation

Multiple-surface segmentation in optical coherence tomography (OCT) images is a challenging problem, further complicated by the frequent presence of weak image boundaries. Recently, many deep learning-based methods have been developed for this task and yield remarkable performance. Unfortunately, due to the scarcity of training data in medical imaging, it is challenging for deep learning networks to learn the global structure of the target surfaces, including surface smoothness. To bridge this gap, this study proposes to seamlessly unify a U-Net for feature learning with a constrained differentiable dynamic programming module to achieve end-to-end learning for retina OCT surface segmentation to explicitly enforce surface smoothness. It effectively utilizes the feedback from the downstream model optimization module to guide feature learning, yielding better enforcement of global structures of the target surfaces. Experiments on Duke AMD (age-related macular degeneration) and JHU MS (multiple sclerosis) OCT data sets for retinal layer segmentation demonstrated that the proposed method was able to achieve subvoxel accuracy on both datasets, with the mean absolute surface distance (MASD) errors of 1.88 ± 1.96μm and 2.75 ± 0.94μm, respectively, over all the segmented surfaces.

Visual Pathway Involvement in NMDA Receptor Encephalitis: A Clinical, Optical Coherence Tomography, and 18-Fluorodeoxyglucose PET/CT Approach

BACKGROUND

Anti-NMDA receptor (NMDAR) encephalitis patients have been reported to exhibit visual dysfunction without retinal thinning. The objective of our study was to examine the involvement of the visual pathway structure and function in anti-NMDAR encephalitis by assessing postrecovery visual function and retinal structure, and acute-phase occipital cortex function.

METHODS

In this cross-sectional study, patients diagnosed with anti-NMDAR encephalitis per consensus criteria underwent postrecovery visual acuity (VA) testing and optical coherence tomography (OCT) with automated retinal layer segmentation. Clinical data and acute-phase brain 18F-fluorodeoxyglucose (FDG) PET/CT (performed within 90 days of symptom onset, assessed qualitatively and semi-quantitatively) were retrospectively analyzed. VA and OCT measures were compared between anti-NMDAR and age, sex, and race-matched healthy controls (HC). When available, FDG-PET/CT metabolism patterns were analyzed for correlations with VA, and OCT measures.

RESULTS

A total of 16 anti-NMDAR (32 eyes) and 32 HC (64 eyes) were included in the study. Anti-NMDAR exhibited lower low-contrast VA (2.5% contrast: -4.4 letters [95% CI; -8.5 to -0.3]; P = 0.04, 1.25% contrast: -6.8 letters [95%CI; -12 to -1.7]; P = 0.01) compared with HC, but no differences were found on OCT-derived retinal layer thicknesses. Acute-phase FDG-PET/CT medial occipital cortex metabolism did not correlate with follow-up low-contrast VA or ganglion cell/inner plexiform layer thickness (GCIPL) (n = 7, 2.5% contrast: r = -0.31; P = 0.50, 1.25% contrast: r = -0.34; P = 0.45, GCIPL: r = -0.04; P = 0.94).

CONCLUSIONS

Although the visual system seems to be involved in anti-NMDAR encephalitis, no retinal structural or occipital cortex functional abnormalities seem to be responsible for the visual dysfunction. When detected acutely, occipital lobe hypometabolism in anti-NMDAR encephalitis does not seem to associate with subsequent retrograde trans-synaptic degenerative phenomena, potentially reflecting reversible neuronal/synaptic dysfunction in the acute phase of the illness rather than neuronal degeneration.

Longitudinal deep network for consistent OCT layer segmentation

Retinal layer thickness is an important bio-marker for people with multiple sclerosis (PwMS). In clinical practice, retinal layer thickness changes in optical coherence tomography (OCT) are widely used for monitoring multiple sclerosis (MS) progression. Recent developments in automated retinal layer segmentation algorithms allow cohort-level retina thinning to be observed in a large study of PwMS. However, variability in these results make it difficult to identify patient-level trends; this prevents patient specific disease monitoring and treatment planning using OCT. Deep learning based retinal layer segmentation algorithms have achieved state-of-the-art accuracy, but the segmentation is performed on each individual scan without utilizing longitudinal information, which can be important in reducing segmentation error and reveal subtle changes in retinal layers. In this paper, we propose a longitudinal OCT segmentation network which achieves more accurate and consistent layer thickness measurements for PwMS.

Mixed multiscale BM4D for three-dimensional optical coherence tomography denoising

A multiscale extension for the well-known block matching and 4D filtering (BM4D) method is proposed by analyzing and extending the wavelet subbands denoising method in such a way that the proposed method avoids directly denoising detail subbands, which considerably simplifies the computations and makes the multiscale processing feasible in 3D. To this end, we first derive the multiscale construction method in 2D and propose multiscale extensions for three 2D natural image denoising methods. Then, the derivation is extended to 3D by proposing mixed multiscale BM4D (mmBM4D) for optical coherence tomography (OCT) image denoising. We tested mmBM4D on three public OCT datasets captured by various imaging devices. The experiments revealed that mmBM4D significantly outperforms its original counterpart and performs on par with the state-of-the-art OCT denoising methods. In terms of peak-signal-to-noise-ratio (PSNR), mmBM4D surpasses the original BM4D by more than 0.68 decibels over the first dataset. In the second and third datasets, significant improvements in the mean to standard deviation ratio, contrast to noise ratio, and equivalent number of looks were achieved. Furthermore, on the downstream task of retinal layer segmentation, the layer quality preservation of the compared OCT denoising methods is evaluated.

Effects of Myopia on Rates of Change in Optical Coherence Tomography Measured Retinal Layer Thicknesses in People with Multiple Sclerosis and Healthy Controls

PURPOSE

To quantify the associations of myopia with longitudinal changes in retinal layer thicknesses in people with multiple sclerosis (PwMS) and healthy controls (HC).

METHODS

A cohort of PwMS and HC with recorded refractive error (RE) prospectively scanned on Cirrus HD-OCT at the Johns Hopkins MS Center was assessed for inclusion. Exclusion criteria included OCT follow-up < 6 months, ocular comorbidities, incidental OCT pathologies, and inadequate scan quality. Eyes were classified as having high myopia (HM) (RE≤ -6 diopters), low myopia (LM) (RE> -6 and ≤ -3 diopters), or no myopia (NM) (RE> -3 and ≤ +2.75). Linear mixed-effects regression models were used in analyses.

RESULTS

A total of 213 PwMS (eyes: 67 HM, 98 LM, 207 NM) and 80 HC (eyes: 26 HM, 37 LM, 93 NM) were included. Baseline average ganglion cell/inner plexiform (GCIPL) and peri-papillary retinal nerve fiber layer (pRNFL) thicknesses were lower in MS HM compared with MS NM (diff: -3.2 µm, 95% CI: -5.5 to -0.8, p = 0.008 and -5.3 µm, 95% CI: -9.0 to -1.7, p = 0.004, respectively), and similarly in HC HM, as compared with HC NM. Baseline superior, inferior, and nasal pRNFL thicknesses were lower in HM compared with NM, while temporal pRNFL thickness was higher, both in MS and HC (MS: 7.1 µm, 95% CI: 2.7-11.6, p = 0.002; HC: 4.7 µm, 95% CI: -0.3 to 9.7, p = 0.07). No longitudinal differences in rates of GCIPL change were noted between HM and LM vs. NM, either in MS or HC.

CONCLUSION

Cross-sectional differences in average GCIPL and pRNFL thicknesses are commonly seen in people with HM as compared to reference normative values from people with NM and can lead to false attribution of pathology if RE is not taken into account. However, our study suggests that longitudinal changes in average GCIPL thickness in PwMS with myopia are similar in magnitude to PwMS with NM, and therefore are appropriate for monitoring disease-related pathology.

Trans-Synaptic Degeneration Following Acute Optic Neuritis in Multiple Sclerosis

OBJECTIVE

To explore longitudinal changes in brain volumetric measures and retinal layer thicknesses following acute optic neuritis (AON) in people with multiple sclerosis (PwMS), to investigate the process of trans-synaptic degeneration, and determine its clinical relevance.

METHODS

PwMS were recruited within 40 days of AON onset (n = 49), and underwent baseline retinal optical coherence tomography and brain magnetic resonance imaging followed by longitudinal tracking for up to 5 years. A comparator cohort of PwMS without a recent episode of AON were similarly tracked (n = 73). Mixed-effects linear regression models were used.

RESULTS

Accelerated atrophy of the occipital gray matter (GM), calcarine GM, and thalamus was seen in the AON cohort, as compared with the non-AON cohort (-0.76% vs -0.22% per year [p = 0.01] for occipital GM, -1.83% vs -0.32% per year [p = 0.008] for calcarine GM, -1.17% vs -0.67% per year [p = 0.02] for thalamus), whereas rates of whole-brain, cortical GM, non-occipital cortical GM atrophy, and T2 lesion accumulation did not differ significantly between the cohorts. In the AON cohort, greater AON-induced reduction in ganglion cell+inner plexiform layer thickness over the first year was associated with faster rates of whole-brain (r = 0.32, p = 0.04), white matter (r = 0.32, p = 0.04), and thalamic (r = 0.36, p = 0.02) atrophy over the study period. Significant relationships were identified between faster atrophy of the subcortical GM and thalamus, with worse visual function outcomes after AON.

INTERPRETATION

These results provide in-vivo evidence for anterograde trans-synaptic degeneration following AON in PwMS, and suggest that trans-synaptic degeneration may be related to clinically-relevant visual outcomes. ANN NEUROL 2023;93:76-87.

High sampling resolution optical coherence tomography reveals potential concurrent reductions in ganglion cell-inner plexiform and inner nuclear layer thickness but not in outer retinal thickness in glaucoma

PURPOSE

To analyse optical coherence tomography (OCT)-derived inner nuclear layer (INL) and outer retinal complex (ORC) measurements relative to ganglion cell-inner plexiform layer (GCIPL) measurements in glaucoma.

METHODS

Glaucoma participants (n = 271) were categorised by 10-2 visual field defect type. Differences in GCIPL, INL and ORC thickness were calculated between glaucoma and matched healthy eyes (n = 548). Hierarchical cluster algorithms were applied to generate topographic patterns of retinal thickness change, with agreement between layers assessed using Cohen's kappa (κ). Differences in GCIPL, INL and ORC thickness within and outside GCIPL regions showing the greatest reductions and Spearman's correlations between layer pairs were compared with 10-2 mean deviation (MD) and pattern standard deviation (PSD) to determine trends with glaucoma severity.

RESULTS

Glaucoma participants with inferior and superior defects presented with concordant GCIPL and INL defects demonstrating mostly fair-to-moderate agreement (κ = 0.145-0.540), which was not observed in eyes with no or ring defects (κ = -0.067-0.230). Correlations (r) with MD and PSD were moderate and weak in GCIPL and INL thickness differences, respectively (GCIPL vs. MD r = 0.479, GCIPL vs. PSD r = -0.583, INL vs. MD r = 0.259, INL vs. PSD r = -0.187, p = <0.0001-0.002), and weak in GCIPL-INL correlations (MD r = 0.175, p = 0.004 and PSD r = 0.154, p = 0.01). No consistent patterns in ORC thickness or correlations were observed.

CONCLUSIONS

In glaucoma, concordant reductions in macular INL and GCIPL thickness can be observed, but reductions in ORC thickness appear unlikely. These findings suggest that trans-synaptic retrograde degeneration may occur in glaucoma and could indicate the usefulness of INL thickness in evaluating glaucomatous damage.

Use of retinal optical coherence tomography to differentiate suspected neuromyelitis optica spectrum disorder from multiple sclerosis: A cross-sectional study

BACKGROUND

Retinal optical coherence tomography (OCT) can differentiate definite NMOSD (dNMOSD) from multiple sclerosis (MS), but has not been evaluated in patients with a high clinical suspicion of NMOSD and not fulfilling the current consensus diagnostic criteria, referred in this paper as "potential" NMOSD (pNMOSD).

AIM

To compare the retinal OCT measurements between patients with pNMOSD, dNMOSD, MS, and reference healthy controls (HC).

MATERIAL AND METHODS

In this cross-sectional study, clinical and demographic characteristics, as well as OCT measurements of peripapillary retinal nerve fiber layer (pRNFL), inner nuclear layer (INL), macular retinal nerve fiber layer (mRNFL), outer nuclear layer (ONL) ganglion cell/inner plexiform layer (GCIPL), and macular volume (MV) were compared between groups. Mixed-effects regression models adjusting for within-patient inter-eye correlations, controlling for age, gender, disease duration and history of optic neuritis per eye were explored. Subgroup analyses were performed on eyes with previous optic neuritis.

RESULTS

234 eyes (20 pNMOSD, 33 dNMOSD, 138 MS, and 43 HC) were included. Controlling for age, gender, disease duration, and history of optic neuritis per eye, pNMOSD eyes showed decreased GCIPL, pRNFL, mRNFL and MV thicknesses, similar to eyes with dNMOSD, but significantly thinner than MS and HC subjects' eyes. Similar results were obtained for the pRNFL, mRNFL, GCIPL, INL and MV thickness in the subgroup analysis exploring only eyes with history of optic neuritis (12 pNMOSD, 15 dNMOSD, and 27 MS).

CONCLUSION

Retinal OCT measurements in patients with pNMOSD were similar to dNMOSD, but significantly lower than patients with MS and healthy controls. This suggests that retinal OCT measures could be helpful markers supportive of NMOSD diagnosis and should be explored in larger studies as a valuable addition to the current consensus diagnostic criteria.

Mitochondrial measures in neuronally enriched extracellular vesicles predict brain and retinal atrophy in multiple sclerosis

BACKGROUND

Mitochondrial dysfunction plays an important role in multiple sclerosis (MS) disease progression. Plasma extracellular vesicles are a potential source of novel biomarkers in MS, and some of these are derived from mitochondria and contain functional mitochondrial components.

OBJECTIVE

To evaluate the relationship between levels of mitochondrial complex IV and V activity in neuronally enriched extracellular vesicles (NEVs) and brain and retinal atrophy as assessed using serial magnetic resonance imaging (MRI) and optical coherence tomography (OCT).

METHODS

Our cohort consisted of 48 people with MS. NEVs were immunocaptured from plasma and mitochondrial complex IV and V activity levels were measured. Subjects underwent OCT every 6 months and brain MRI annually. The associations between baseline mitochondrial complex IV and V activities and brain substructure and retinal thickness changes were estimated utilizing linear mixed-effects models.

RESULTS

We found that higher mitochondrial complex IV activity and lower mitochondrial complex V activity levels were significantly associated with faster whole-brain volume atrophy. Similar results were found with other brain substructures and retinal layer atrophy.

CONCLUSION

Our results suggest that mitochondrial measures in circulating NEVs could serve as potential biomarkers of disease progression and provide the rationale for larger follow-up longitudinal studies.

Structure-function correlates of vision loss in neuromyelitis optica spectrum disorders

Optic neuritis (ON) in neuromyelitis optica spectrum disorders (NMOSD) regularly leads to more profound vision loss compared to multiple sclerosis (MS) and myelin-oligodendrocyte-glycoprotein-antibody associated disease (MOGAD). Here we investigate ON-related vision loss in NMOSD compared to MS and MOGAD in order to identify neuroaxonal and retinal contributors to visual dysfunction. In this retrospective study we included patients with aquaporin-4-antibody seropositive NMOSD (n = 28), MOGAD (n = 14), MS (n = 29) and controls (n = 14). We assessed optic nerve damage and fovea morphometry by optical coherence tomography. Visual function was assessed as high (HCVA) and low contrast visual acuity (LCVA), and visual fields' mean deviation (MD). In all diseases, lower visual function was associated with peripapillary retinal nerve fiber layer (pRNFL) and ganglion cell and inner plexiform layer (GCIP) thinning following a broken stick model, with pRNFL and GCIP cutoff point at ca. 60 µm. HCVA loss per µm pRNFL and GCIP thinning was stronger in NMOSD compared with MOGAD. Foveal inner rim volume contributed to MD and LCVA in NMOSD eyes, only. Together these data supports that visual dysfunction in NMOSD is associated with neuroaxonal damage beyond the effect seen in MS and MOGAD. A primary retinopathy, respectively Müller cell pathology, may contribute to this effect.

Rate of Retinal Layer Thinning as a Biomarker for Conversion to Progressive Disease in Multiple Sclerosis

Background and Objectives

The diagnosis of secondary progressive multiple sclerosis (SPMS) is often delayed because of the lack of objective clinical tools, which increases the diagnostic uncertainty and hampers the therapeutic development in progressive multiple sclerosis (MS). Optical coherence tomography (OCT) has been proposed as a promising biomarker of progressive neurodegeneration. To explore longitudinal changes in the thicknesses of retinal layers on OCT in individuals with relapsing-remitting MS (RRMS) who converted to SPMS vs matched patients with RRMS who did not convert to SPMS. Our hypothesis is that the 2 cohorts exhibit different rates of retinal thinning.

Methods

From our prospective observational cohort of patients with MS at the American University of Beirut, we selected patients with RRMS who converted to SPMS during the observation period and patients with RRMS, matched by age, disease duration, and Expanded Disability Status Scale (EDSS) at the first visit. Baseline retinal measurements were obtained using spectral domain OCT, and all patients underwent clinical and OCT evaluation every 6–12 months on average throughout the study period (mean = 4 years). Mixed-effect regression models were used to assess the annualized rates of retinal changes and the differences between the 2 groups and between converters to SPMS before and after their conversion.

Results

A total of 61 participants were selected (21 SPMS and 40 RRMS). There were no differences in baseline characteristics and retinal measurements between the 2 groups. The annualized rates of thinning of all retinal layers, except for macular volume, were greater in converters before conversion compared with nonconverters by 112% for peripapillary retinal nerve fiber layer (p = 0.008), 344% for tRNFL (p < 0.0001), and 82% for cell-inner plexiform layer (GCIPL) (p = 0.002). When comparing the annualized rate of thinning for the same patients with SPMS before and after conversion, no significant differences were found except for tRNFL and GCIPL with slower thinning rates postconversion (46% and 68%, respectively).

Discussion

Patients who converted to SPMS exhibited faster retinal thinning as reflected on OCT. Longitudinal assessment of retinal thinning could confirm the transition to SPMS and help with the therapeutic decision making for patients with MS with clinical suspicion of disease progression.

Automatic Segmentation and Measurement of Choroid Layer in High Myopia for OCT Imaging Using Deep Learning

Automatic segmentation and measurement of the choroid layer is useful in studying of related fundus diseases, such as diabetic retinopathy and high myopia. However, most algorithms are not helpful for choroid layer segmentation due to its blurred boundaries and complex gradients. Therefore, this paper aimed to propose a novel choroid segmentation method that combines image enhancement and attention-based dense (AD) U-Net network. The choroidal images obtained from optical coherence tomography (OCT) are pre-enhanced by algorithms that include flattening, filtering, and exponential and linear enhancement to reduce choroid-independent information. Experimental results obtained from 800 OCT B-scans of the choroid layers from both normal eyes and high myopia showed that image enhancement significantly increased the performance of ADU-Net, with an AUC of 99.51% and a DSC of 97.91%. The accuracy of segmentation using the ADU-Net method with image enhancement is superior to that of the existing networks. In addition, we describe some algorithms that can measure automatically choroidal foveal thickness and the volume of adjacent areas. Statistical analyses of the choroidal parameters variation indicated that compared with normal eyes, high myopia has a reduction of 86.3% of the choroidal foveal thickness and 90% of the adjacent volume. It proved that high myopia is likely to cause choroid layer attenuation. These algorithms would have wide application in the diagnosis and precaution of related fundus lesions caused by choroid thinning from high myopia in future studies.

Deep learning: applications in retinal and optic nerve diseases

Deep learning (DL) represents a paradigm-shifting, burgeoning field of research with emerging clinical applications in optometry. Unlike traditional programming, which relies on human-set specific rules, DL works by exposing the algorithm to a large amount of annotated data and allowing the software to develop its own set of rules (i.e. learn) by adjusting the parameters inside the model (network) during a training process in order to complete the task on its own. One major limitation of traditional programming is that, with complex tasks, it may require an extensive set of rules to accurately complete the assignment. Additionally, traditional programming can be susceptible to human bias from programmer experience. With the dramatic increase in the amount and the complexity of clinical data, DL has been utilised to automate data analysis and thus to assist clinicians in patient management. This review will present the latest advances in DL, for managing posterior eye diseases as well as DL-based solutions for patients with vision loss.

Association of Serum Neurofilament Light Chain With Inner Retinal Layer Thinning in Multiple Sclerosis

BACKGROUND AND OBJECTIVES

Serum neurofilament light chain (sNfL) and optical coherence tomography (OCT)-derived retinal measures (including peripapillary retinal nerve fiber layer [pRNFL] and macular ganglion cell layer/inner plexiform layer [GCIPL] thickness) have been proposed as biomarkers of neurodegeneration in multiple sclerosis (MS). However, studies evaluating the associations between sNfL and OCT-derived retinal measures in MS are limited.

METHODS

In this retrospective analysis of a longitudinal, observational, single-center cohort study, sNfL levels were measured in people with MS and healthy controls (HCs) using single molecule array. Participants with MS were followed with serial OCT for a median follow-up of 4.5 years. Eyes with optic neuritis (ON) within 6 months of baseline OCT or ON during follow-up were excluded. Age-normative cutoffs of sNfL were derived using the HC data, and MS participants with sNfL greater than the 97.5th percentile for age were classified as having elevated sNfL (sNfL-E). Analyses were performed with mixed-effects linear regression models and adjusted for age, sex, race, and history of ON.

RESULTS

A total of 130 HCs (age: 42.4 ± 14.2 years; 62% female) and 403 people with MS (age: 43.1 ± 12.0 years; 78% female) were included. Elevated sNfL levels were present at baseline in 80 participants with MS (19.9%). At baseline, sNfL-E participants had modestly lower pRNFL (-3.03 ± 1.50 μm; p = 0.044) and GCIPL thickness (-2.74 ± 1.02 μm; p = 0.007). As compared with those with sNfL within the reference range, eyes from NfL-E participants exhibited faster longitudinal thinning of the pRNFL (45% faster; -0.74 vs -0.51 μm/y; p = 0.015) and GCIPL (25% faster; -0.35 vs -0.28 μm/y; p = 0.021). Significant differences in rates of pRNFL and GCIPL thinning between sNfL groups were found only in those with relapsing-remitting MS but not progressive MS.

DISCUSSION

Elevated baseline sNfL is associated with accelerated rates of retinal neuroaxonal loss in relapsing-remitting MS, independent of overt ON, but may be less reflective of retinal neurodegeneration in progressive MS.

A single-step regression method based on transformer for retinal layer segmentation

The shape and structure of retinal layers are basic characteristics for the diagnosis of many ophthalmological diseases. Based on B-Scans of optical coherence tomography, most of retinal layer segmentation methods are composed of two-steps: classifying pixels and extracting retinal layers, in which the optimization of two independent steps decreases the accuracy. Although the methods based on deep learning are highly accurate, they require a large amount of labeled data. This paper proposes a single-step method based on transformer for retinal layer segmentation, which is trained by axial data (A-Scans), to obtain the boundary of each layer. The proposed method was evaluated on two public data sets. The first one contains eight retinal layer boundaries for diabetic macular edema, and the second one contains nine retinal layer boundaries for healthy controls and subjects with multiple sclerosis. Its absolute average distance errors are 0.99 pixels and 3.67 pixels, respectively, for the two sets, and its root mean square error is 1.29 pixels for the latter set. In addition, its accuracy is acceptable even if the training data is reduced to 0.3. The proposed method achieves state-of-the-art performance while maintaining the correct topology and requires less labeled data.

Automatic Segmentation of the Optic Nerve Head Region in Optical Coherence Tomography: A Methodological Review

The optic nerve head (ONH) represents the intraocular section of the optic nerve, which is prone to damage by intraocular pressure (IOP). The advent of optical coherence tomography (OCT) has enabled the evaluation of novel ONH parameters, namely the depth and curvature of the lamina cribrosa (LC). Together with the Bruch's membrane minimum-rim-width (BMO-MRW), these seem to be promising ONH parameters for diagnosis and monitoring of retinal diseases such as glaucoma. Nonetheless, these OCT derived biomarkers are mostly extracted through manual segmentation, which is time-consuming and prone to bias, thus limiting their usability in clinical practice. The automatic segmentation of ONH in OCT scans could further improve the current clinical management of glaucoma and other diseases. This review summarizes the current state-of-the-art in automatic segmentation of the ONH in OCT. PubMed and Scopus were used to perform a systematic review. Additional works from other databases (IEEE, Google Scholar and ARVO IOVS) were also included, resulting in a total of 29 reviewed studies. For each algorithm, the methods, the size and type of dataset used for validation, and the respective results were carefully analysed. The results show a lack of consensus regarding the definition of segmented regions, extracted parameters and validation approaches, highlighting the importance and need of standardized methodologies for ONH segmentation. Only with a concrete set of guidelines, these automatic segmentation algorithms will build trust in data-driven segmentation models and be able to enter clinical practice.

Retinal layer segmentation in optical coherence tomography (OCT) using a 3D deep-convolutional regression network for patients with age-related macular degeneration

Introduction - Retinal layer segmentation in optical coherence tomography (OCT) images is an important approach for detecting and prognosing disease. Automating segmentation using robust machine learning techniques lead to computationally efficient solutions and significantly reduces the cost of labor-intensive labeling, which is traditionally performed by trained graders at a reading center, sometimes aided by semi-automated algorithms. Although several algorithms have been proposed since the revival of deep learning, eyes with severe pathological conditions continue to challenge fully automated segmentation approaches. There remains an opportunity to leverage the underlying spatial correlations between the retinal surfaces in the segmentation approach. Methods - Some of these proposed traditional methods can be expanded to utilize the three-dimensional spatial context governing the retinal image volumes by replacing the use of 2D filters with 3D filters. Towards this purpose, we propose a spatial-context, continuity and anatomical relationship preserving semantic segmentation algorithm, which utilizes the 3D spatial context from the image volumes with the use of 3D filters. We propose a 3D deep neural network capable of learning the surface positions of the layers in the retinal volumes. Results - We utilize a dataset of OCT images from patients with Age-related Macular Degeneration (AMD) to assess performance of our model and provide both qualitative (including segmentation maps and thickness maps) and quantitative (including error metric comparisons and volumetric comparisons) results, which demonstrate that our proposed method performs favorably even for eyes with pathological changes caused by severe retinal diseases. The Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) for patients with a wide range of AMD severity scores (0-11) were within 0.84±0.41 and 1.33±0.73 pixels, respectively, which are significantly better than some of the other state-of-the-art algorithms. Conclusion - The results demonstrate the utility of extracting features from the entire OCT volume by treating the volume as a correlated entity and show the benefit of utilizing 3D autoencoder based regression networks for smoothing the approximated retinal layers by inducing shape based regularization constraints.

Intraretinal Layer Segmentation Using Cascaded Compressed U-Nets

Reliable biomarkers quantifying neurodegeneration and neuroinflammation in central nervous system disorders such as Multiple Sclerosis, Alzheimer's dementia or Parkinson's disease are an unmet clinical need. Intraretinal layer thicknesses on macular optical coherence tomography (OCT) images are promising noninvasive biomarkers querying neuroretinal structures with near cellular resolution. However, changes are typically subtle, while tissue gradients can be weak, making intraretinal segmentation a challenging task. A robust and efficient method that requires no or minimal manual correction is an unmet need to foster reliable and reproducible research as well as clinical application. Here, we propose and validate a cascaded two-stage network for intraretinal layer segmentation, with both networks being compressed versions of U-Net (CCU-INSEG). The first network is responsible for retinal tissue segmentation from OCT B-scans. The second network segments eight intraretinal layers with high fidelity. At the post-processing stage, we introduce Laplacian-based outlier detection with layer surface hole filling by adaptive non-linear interpolation. Additionally, we propose a weighted version of focal loss to minimize the foreground-background pixel imbalance in the training data. We train our method using 17,458 B-scans from patients with autoimmune optic neuropathies, i.e., multiple sclerosis, and healthy controls. Voxel-wise comparison against manual segmentation produces a mean absolute error of 2.3 μm, outperforming current state-of-the-art methods on the same data set. Voxel-wise comparison against external glaucoma data leads to a mean absolute error of 2.6 μm when using the same gold standard segmentation approach, and 3.7 μm mean absolute error in an externally segmented data set. In scans from patients with severe optic atrophy, 3.5% of B-scan segmentation results were rejected by an experienced grader, whereas this was the case in 41.4% of B-scans segmented with a graph-based reference method. The validation results suggest that the proposed method can robustly segment macular scans from eyes with even severe neuroretinal changes.

Diagnosis of Retinal Diseases Based on Bayesian Optimization Deep Learning Network Using Optical Coherence Tomography Images

Retinal abnormalities have emerged as a serious public health concern in recent years and can manifest gradually and without warning. These diseases can affect any part of the retina, causing vision impairment and indeed blindness in extreme cases. This necessitates the development of automated approaches to detect retinal diseases more precisely and, preferably, earlier. In this paper, we examine transfer learning of pretrained convolutional neural network (CNN) and then transfer it to detect retinal problems from Optical Coherence Tomography (OCT) images. In this study, pretrained CNN models, namely, VGG16, DenseNet201, InceptionV3, and Xception, are used to classify seven different retinal diseases from a dataset of images with and without retinal diseases. In addition, to choose optimum values for hyperparameters, Bayesian optimization is applied, and image augmentation is used to increase the generalization capabilities of the developed models. This research also provides a comparison of the proposed models as well as an analysis of them. The accuracy achieved using DenseNet201 on the Retinal OCT Image dataset is more than 99% and offers a good level of accuracy in classifying retinal diseases compared to other approaches, which only detect a small number of retinal diseases.

Clinical Evaluations of Macular Structure-Function Concordance With and Without Drasdo Displacement

Purpose

The purpose of this study was to compare concordance between ganglion cell-inner plexiform layer (GCIPL) data from the Cirrus optical coherence tomographer (OCT) Ganglion Cell Analysis (GCA) and visual fields (VFs), with and without Drasdo displacement.

Methods

From 296 open-angle glaucoma participants, GCIPL deviation and raw thickness data were extracted over locations per the 10-2 VF test grid, with and without application of Drasdo displacement, with global and eccentricity-dependent sensitivities and specificities calculated for both. With OCT and VF data classified as within or outside normative limits, pattern deviation values were compared using paired t-tests and Spearman correlations. Regression models were applied to pattern deviation values as a function of GCIPL thickness, and differences in model performance with and without displacement were compared using extra sums-of-squares F tests.

Results

There were small but significant improvements in global specificity without displacement (0.58-0.59 with displacement and 0.61 without displacement), without notable differences in sensitivity (0.77-0.78 with displacement and 0.76-0.78 without displacement). At abnormal VF locations and without displacement, a higher proportion of correct OCT classifications (P = 0.0008) and significant correlation with worsening pattern deviation values were observed (r = 0.50, P = 0.002). Regression models indicated significantly steeper slopes with Drasdo displacement centrally (P = 0.002-0.04).

Conclusions

With GCA deviation maps, small improvements in structure-function concordance were observed without displacement, which are unlikely to be clinically meaningful. Using GCIPL thickness data, significantly better structure-function concordance was observed centrally with Drasdo displacement.

Translational Relevance

Applying Drasdo displacement on probability-based reports is unlikely to alter clinical impressions of structure-function concordance, but applying displacement with GCIPL thickness data may improve detection of structure-function concordance.

Automated retinal boundary segmentation of optical coherence tomography images using an improved Canny operator

Retinal segmentation is a prerequisite for quantifying retinal structural features and diagnosing related ophthalmic diseases. Canny operator is recognized as the best boundary detection operator so far, and is often used to obtain the initial boundary of the retina in retinal segmentation. However, the traditional Canny operator is susceptible to vascular shadows, vitreous artifacts, or noise interference in retinal segmentation, causing serious misdetection or missed detection. This paper proposed an improved Canny operator for automatic segmentation of retinal boundaries. The improved algorithm solves the problems of the traditional Canny operator by adding a multi-point boundary search step on the basis of the original method, and adjusts the convolution kernel. The algorithm was used to segment the retinal images of healthy subjects and age-related macular degeneration (AMD) patients; eleven retinal boundaries were identified and compared with the results of manual segmentation by the ophthalmologists. The average difference between the automatic and manual methods is: 2–6 microns (1–2 pixels) for healthy subjects and 3–10 microns (1–3 pixels) for AMD patients. Qualitative method is also used to verify the accuracy and stability of the algorithm. The percentage of “perfect segmentation” and “good segmentation” is 98% in healthy subjects and 94% in AMD patients. This algorithm can be used alone or in combination with other methods as an initial boundary detection algorithm. It is easy to understand and improve, and may become a useful tool for analyzing and diagnosing eye diseases.

Globally optimal OCT surface segmentation using a constrained IPM optimization

Segmentation of multiple surfaces in optical coherence tomography (OCT) images is a challenging problem, further complicated by the frequent presence of weak boundaries, varying layer thicknesses, and mutual influence between adjacent surfaces. The traditional graph-based optimal surface segmentation method has proven its effectiveness with its ability to capture various surface priors in a uniform graph model. However, its efficacy heavily relies on handcrafted features that are used to define the surface cost for the "goodness" of a surface. Recently, deep learning (DL) is emerging as a powerful tool for medical image segmentation thanks to its superior feature learning capability. Unfortunately, due to the scarcity of training data in medical imaging, it is nontrivial for DL networks to implicitly learn the global structure of the target surfaces, including surface interactions. This study proposes to parameterize the surface cost functions in the graph model and leverage DL to learn those parameters. The multiple optimal surfaces are then simultaneously detected by minimizing the total surface cost while explicitly enforcing the mutual surface interaction constraints. The optimization problem is solved by the primal-dual interior-point method (IPM), which can be implemented by a layer of neural networks, enabling efficient end-to-end training of the whole network. Experiments on spectral-domain optical coherence tomography (SD-OCT) retinal layer segmentation demonstrated promising segmentation results with sub-pixel accuracy.

Multilevel Deep Feature Generation Framework for Automated Detection of Retinal Abnormalities Using OCT Images

Optical coherence tomography (OCT) images coupled with many learning techniques have been developed to diagnose retinal disorders. This work aims to develop a novel framework for extracting deep features from 18 pre-trained convolutional neural networks (CNN) and to attain high performance using OCT images. In this work, we have developed a new framework for automated detection of retinal disorders using transfer learning. This model consists of three phases: deep fused and multilevel feature extraction, using 18 pre-trained networks and tent maximal pooling, feature selection with ReliefF, and classification using the optimized classifier. The novelty of this proposed framework is the feature generation using widely used CNNs and to select the most suitable features for classification. The extracted features using our proposed intelligent feature extractor are fed to iterative ReliefF (IRF) to automatically select the best feature vector. The quadratic support vector machine (QSVM) is utilized as a classifier in this work. We have developed our model using two public OCT image datasets, and they are named database 1 (DB1) and database 2 (DB2). The proposed framework can attain 97.40% and 100% classification accuracies using the two OCT datasets, DB1 and DB2, respectively. These results illustrate the success of our model.

Assessing the validity of a cross-platform retinal image segmentation tool in normal and diseased retina

Comparing automated retinal layer segmentation using proprietary software (Heidelberg Spectralis HRA + OCT) and cross-platform Optical Coherence Tomography (OCT) segmentation software (Orion). Image segmentations of normal and diseased (iAMD, DME) eyes were performed using both softwares and then compared to the 'gold standard' of manual segmentation. A qualitative assessment and quantitative (layer volume) comparison of segmentations were performed. Segmented images from the two softwares were graded by two masked graders and in cases with difference, a senior retina specialist made a final independent decisive grading. Cross-platform software was significantly better than the proprietary software in the segmentation of NFL and INL layers in Normal eyes. It generated significantly better segmentation only for NFL in iAMD and for INL and OPL layers in DME eyes. In normal eyes, all retinal layer volumes calculated by the two softwares were moderate-strongly correlated except OUTLY. In iAMD eyes, GCIPL, INL, ONL, INLY, TRV layer volumes were moderate-strongly correlated between softwares. In eyes with DME, all layer volume values were moderate-strongly correlated between softwares. Cross-platform software can be used reliably in research settings to study the retinal layers as it compares well against manual segmentation and the commonly used proprietary software for both normal and diseased eyes.

LF-UNet - A novel anatomical-aware dual-branch cascaded deep neural network for segmentation of retinal layers and fluid from optical coherence tomography images

Computer-assistant diagnosis of retinal disease relies heavily on the accurate detection of retinal boundaries and other pathological features such as fluid accumulation. Optical coherence tomography (OCT) is a non-invasive ophthalmological imaging technique that has become a standard modality in the field due to its ability to detect cross-sectional retinal pathologies at the micrometer level. In this work, we presented a novel framework to achieve simultaneous retinal layers and fluid segmentation. A dual-branch deep neural network, termed LF-UNet, was proposed which combines the expansion path of the U-Net and original fully convolutional network, with a dilated network. In addition, we introduced a cascaded network framework to include the anatomical awareness embedded in the volumetric image. Cross validation experiments showed that the proposed LF-UNet has superior performance compared to the state-of-the-art methods, and that incorporating the relative positional map structural prior information could further improve the performance regardless of the network. The generalizability of the proposed network was demonstrated on an independent dataset acquired from the same types of device with different field of view, or images acquired from different device.

Multi-omic evaluation of metabolic alterations in multiple sclerosis identifies shifts in aromatic amino acid metabolism

The circulating metabolome provides unique insights into multiple sclerosis (MS) pathophysiology, but existing studies are relatively small or characterized limited metabolites. We test for differences in the metabolome between people with MS (PwMS; n = 637 samples) and healthy controls (HC; n = 317 samples) and assess the association between metabolomic profiles and disability in PwMS. We then assess whether metabolic differences correlate with changes in cellular gene expression using publicly available scRNA-seq data and whether identified metabolites affect human immune cell function. In PwMS, we identify striking abnormalities in aromatic amino acid (AAA) metabolites (p = 2.77E-18) that are also strongly associated with disability (p = 1.01E-4). Analysis of scRNA-seq data demonstrates altered AAA metabolism in CSF and blood-derived monocyte cell populations in PwMS. Treatment with AAA-derived metabolites in vitro alters monocytic endocytosis and pro-inflammatory cytokine production. We identify shifts in AAA metabolism resulting in the reduced production of immunomodulatory metabolites and increased production of metabotoxins in PwMS.

Retinal Optical Coherence Tomography in Neuromyelitis Optica

Background and Objectives

To determine optic nerve and retinal damage in aquaporin-4 antibody (AQP4-IgG)-seropositive neuromyelitis optica spectrum disorders (NMOSD) in a large international cohort after previous studies have been limited by small and heterogeneous cohorts.

Methods

The cross-sectional Collaborative Retrospective Study on retinal optical coherence tomography (OCT) in neuromyelitis optica collected retrospective data from 22 centers. Of 653 screened participants, we included 283 AQP4-IgG–seropositive patients with NMOSD and 72 healthy controls (HCs). Participants underwent OCT with central reading including quality control and intraretinal segmentation. The primary outcome was thickness of combined ganglion cell and inner plexiform (GCIP) layer; secondary outcomes were thickness of peripapillary retinal nerve fiber layer (pRNFL) and visual acuity (VA).

Results

Eyes with ON (NMOSD-ON, N = 260) or without ON (NMOSD-NON, N = 241) were assessed compared with HCs (N = 136). In NMOSD-ON, GCIP layer (57.4 ± 12.2 μm) was reduced compared with HC (GCIP layer: 81.4 ± 5.7 μm, p < 0.001). GCIP layer loss (−22.7 μm) after the first ON was higher than after the next (−3.5 μm) and subsequent episodes. pRNFL observations were similar. NMOSD-NON exhibited reduced GCIP layer but not pRNFL compared with HC. VA was greatly reduced in NMOSD-ON compared with HC eyes, but did not differ between NMOSD-NON and HC.

Discussion

Our results emphasize that attack prevention is key to avoid severe neuroaxonal damage and vision loss caused by ON in NMOSD. Therapies ameliorating attack-related damage, especially during a first attack, are an unmet clinical need. Mild signs of neuroaxonal changes without apparent vision loss in ON-unaffected eyes might be solely due to contralateral ON attacks and do not suggest clinically relevant progression but need further investigation.

Deep Relation Transformer for Diagnosing Glaucoma With Optical Coherence Tomography and Visual Field Function

Glaucoma is the leading reason for irreversible blindness. Early detection and timely treatment of glaucoma are essential for preventing visual field loss or even blindness. In clinical practice, Optical Coherence Tomography (OCT) and Visual Field (VF) exams are two widely-used and complementary techniques for diagnosing glaucoma. OCT provides quantitative measurements of the optic nerve head (ONH) structure, while VF test is the functional assessment of peripheral vision. In this paper, we propose a Deep Relation Transformer (DRT) to perform glaucoma diagnosis with OCT and VF information combined. A novel deep reasoning mechanism is proposed to explore implicit pairwise relations between OCT and VF information in global and regional manners. With the pairwise relations, a carefully-designed deep transformer mechanism is developed to enhance the representation with complementary information for each modal. Based on reasoning and transformer mechanisms, three successive modules are designed to extract and collect valuable information for glaucoma diagnosis, the global relation module, the guided regional relation module, and the interaction transformer module, namely. Moreover, we build a large dataset, namely ZOC-OCT&VF dataset, which includes 1395 OCT-VF pairs for developing and evaluating our DRT. We conduct extensive experiments to validate the effectiveness of the proposed method. Experimental results show that our method achieves 88.3% accuracy and outperforms the existing single-modal approaches with a large margin. The codes and dataset will be publicly available in the future.

Boundary Aware U-Net for Retinal Layers Segmentation in Optical Coherence Tomography Images

Retinal layers segmentation in optical coherence tomography (OCT) images is a critical step in the diagnosis of numerous ocular diseases. Automatic layers segmentation requires separating each individual layer instance with accurate boundary detection, but remains a challenging task since it suffers from speckle noise, intensity inhomogeneity, and the low contrast around boundary. In this work, we proposed a boundary aware U-Net (BAU-Net) for retinal layers segmentation by detecting accurate boundary. Based on encoder-decoder architecture, we design a dual tasks framework with low-level outputs for boundary detection and high-level outputs for layers segmentation. Specifically, we first use the multi-scale input strategy to enrich the spatial information in the deep features of encoder. For low-level features from encoder, we design an edge aware (EA) module in skip connection to extract the pure edge features. Then, a U-structure feature enhanced (UFE) module is designed in all skip connections to enlarge the features receptive fields from the encoder. Besides, a canny edge fusion (CEF) module is introduced to aforementioned architecture, which can fuse the priory edge information from segmentation task to boundary detection branch for a better predication. Furthermore, we model each boundary as a vertical coordinates distribution for boundary detection. Based on this distribution, a topology guarantee loss with combined A-scan regression loss and structure loss is proposed to make an accurate and guaranteed topological boundary set. The method is evaluated on two public datasets and the results demonstrate that the BAU-Net achieves promising performance than other state-of-the-art methods.

Deep Learning-Based Retinal Nerve Fiber Layer Thickness Measurement of Murine Eyes

Purpose

To design a robust and automated estimation method for measuring the retinal nerve fiber layer (RNFL) thickness using spectral domain optical coherence tomography (SD-OCT).

Methods

We developed a deep learning-based image segmentation network for automated segmentation of the RNFL in SD-OCT B-scans of mouse eyes. In total, 5500 SD-OCT B-scans (5200 B-scans were used as training data with the remaining 300 B-scans used as testing data) were used to develop this segmentation network. Postprocessing operations were then applied on the segmentation results to fill any discontinuities or remove any speckles in the RNFL. Subsequently, a three-dimensional retina thickness map was generated by z-stacking 100 segmentation processed thickness B-scan images together. Finally, the average absolute difference between algorithm predicted RNFL thickness compared to the ground truth manual human segmentation was calculated.

Results

The proposed method achieves an average dice similarity coefficient of 0.929 in the SD-OCT segmentation task and an average absolute difference of 0.0009 mm in thickness estimation task on the basis of the testing dataset. We also evaluated our segmentation algorithm on another biological dataset with SD-OCT volumes for RNFL thickness after the optic nerve crush injury. Results were shown to be comparable between the predicted and manually measured retina thickness values.

Conclusions

Experimental results demonstrate that our automated segmentation algorithm reliably predicts the RNFL thickness in SD-OCT volumes of mouse eyes compared to laborious and more subjective manual SD-OCT RNFL segmentation.

Translational Relevance

Automated segmentation using a deep learning-based algorithm for murine eye OCT effectively and rapidly produced nerve fiber layer thicknesses comparable to manual segmentation.

V, Josephraj AN, E. S. Diagnosis of retinal disorders from Optical Coherence Tomography images using CNN

An efficient automatic decision support system for detection of retinal disorders is important and is the need of the hour. Optical Coherence Tomography (OCT) is the current imaging modality for the early detection of retinal disorders non-invasively. In this work, a Convolution Neural Network (CNN) model is proposed to classify three types of retinal disorders namely: Choroidal neovascularization (CNV), Drusen macular degeneration (DMD) and Diabetic macular edema (DME). The hyperparameters of the model like batch size, number of epochs, dropout rate, and the type of optimizer are tuned using random search optimization method for better performance to classify different retinal disorders. The proposed architecture provides an accuracy of 97.01%, sensitivity of 93.43%, and 98.07% specificity and it outperformed other existing models, when compared. The proposed model can be used for the large-scale screening of retinal disorders effectively.

Foveal Remodeling of Retinal Microvasculature in Parkinson's Disease

Background

Retinal microvascular alterations have been previously described in Parkinson’s disease (PD) patients using optical coherence tomography angiography (OCT-A). However, an extensive description of retinal vascular morphological features, their association with PD-related clinical variables and their potential use as diagnostic biomarkers has not been explored.

Methods

We performed a cross-sectional study including 49 PD patients (87 eyes) and 40 controls (73 eyes). Retinal microvasculature was evaluated with Spectralis OCT-A and cognitive status with Montreal Cognitive Assessment. Unified PD Rating Scale and disease duration were recorded in patients. We extracted microvascular parameters from superficial and deep vascular plexuses of the macula, including the area and circularity of foveal avascular zone (FAZ), skeleton density, perfusion density, vessel perimeter index, vessel mean diameter, fractal dimension (FD) and lacunarity using Python and MATLAB. We compared the microvascular parameters between groups and explored their association with thickness of macular layers and clinical outcomes. Data were analyzed with General Estimating Equations (GEE) and adjusted for age, sex, and hypertension. Logistic regression GEE models were fitted to predict diagnosis of PD versus controls from microvascular, demographic, and clinical data. The discrimination ability of models was tested with receiver operating characteristic curves.

Results

FAZ area was significantly smaller in patients compared to controls in superficial and deep plexuses, whereas perfusion density, skeleton density, FD and lacunarity of capillaries were increased in the foveal zone of PD. In the parafovea, microvascular parameters of superficial plexus were associated with ganglion cell-inner plexiform layer thickness, but this was mainly driven by PD with mild cognitive impairment. No such associations were observed in controls. FAZ area was negatively associated with cognition in PD (non-adjusted models). Foveal lacunarity, combined with demographic and clinical confounding factors, yielded an outstanding diagnostic accuracy for discriminating PD patients from controls.

Conclusion

Parkinson’s disease patients displayed foveal microvascular alterations causing an enlargement of the vascular bed surrounding FAZ. Parafoveal microvascular alterations were less pronounced but were related to inner retinal layer thinning. Retinal microvascular abnormalities helped discriminating PD from controls. All this supports OCT-A as a potential non-invasive biomarker to reveal vascular pathophysiology and improve diagnostic accuracy in PD.

Quantitative Assessment of Experimental Ocular Inflammatory Disease

Ocular inflammation imposes a high medical burden on patients and substantial costs on the health-care systems that mange these often chronic and debilitating diseases. Many clinical phenotypes are recognized and classifying the severity of inflammation in an eye with uveitis is an ongoing challenge. With the widespread application of optical coherence tomography in the clinic has come the impetus for more robust methods to compare disease between different patients and different treatment centers. Models can recapitulate many of the features seen in the clinic, but until recently the quality of imaging available has lagged that applied in humans. In the model experimental autoimmune uveitis (EAU), we highlight three linked clinical states that produce retinal vulnerability to inflammation, all different from healthy tissue, but distinct from each other. Deploying longitudinal, multimodal imaging approaches can be coupled to analysis in the tissue of changes in architecture, cell content and function. This can enrich our understanding of pathology, increase the sensitivity with which the impacts of therapeutic interventions are assessed and address questions of tissue regeneration and repair. Modern image processing, including the application of artificial intelligence, in the context of such models of disease can lay a foundation for new approaches to monitoring tissue health.

Association of retinal atrophy with cortical lesions and leptomeningeal enhancement in multiple sclerosis on 7T MRI

BACKGROUND

Retinal atrophy in multiple sclerosis (MS) as measured by optical coherence tomography (OCT) correlates with demyelinating lesions and brain atrophy, but its relationship with cortical lesions (CLs) and meningeal inflammation is not well known.

OBJECTIVES

To evaluate the relationship of retinal layer atrophy with leptomeningeal enhancement (LME) and CLs in MS as visualized on 7 Tesla (7T) magnetic resonance imaging (MRI).

METHODS

Forty participants with MS underwent 7T MRI of the brain and OCT. Partial correlation and mixed-effects regression evaluated relationships between MRI and OCT findings.

RESULTS

All participants had CLs and 32 (80%) participants had LME on post-contrast MRI. Ganglion cell/inner plexiform layer (GCIPL) thickness correlated with total CL volume (r =-0.45, p < 0.01). Participants with LME at baseline had thinner macular retinal nerve fiber layer (mRNFL; p = 0.01) and GCIPL (p < 0.01). Atrophy in various retinal layers was faster in those with certain patterns of LME. For example, mRNFL declined -1.113 (-1.974, -0.252) μm/year faster in those with spread/fill-pattern LME foci at baseline compared with those without (p = 0.01).

CONCLUSION

This study associates MRI findings of LME and cortical pathology with thinning of retinal layers as measured by OCT, suggesting a common link between meningeal inflammation, CLs, and retinal atrophy in MS.

Foveal Pit Morphology Characterization: A Quantitative Analysis of the Key Methodological Steps

Disentangling the cellular anatomy that gives rise to human visual perception is one of the main challenges of ophthalmology. Of particular interest is the foveal pit, a concave depression located at the center of the retina that captures light from the gaze center. In recent years, there has been a growing interest in studying the morphology of the foveal pit by extracting geometrical features from optical coherence tomography (OCT) images. Despite this, research has devoted little attention to comparing existing approaches for two key methodological steps: the location of the foveal center and the mathematical modelling of the foveal pit. Building upon a dataset of 185 healthy subjects imaged twice, in the present paper the image alignment accuracy of four different foveal center location methods is studied in the first place. Secondly, state-of-the-art foveal pit mathematical models are compared in terms of fitting error, repeatability, and bias. The results indicate the importance of using a robust foveal center location method to align images. Moreover, we show that foveal pit models can improve the agreement between different acquisition protocols. Nevertheless, they can also introduce important biases in the parameter estimates that should be considered.

Modulation of Retinal Atrophy With Rituximab in Multiple Sclerosis

OBJECTIVE

To investigate the effects of rituximab on retinal atrophy in patients with relapsing-remitting multiple sclerosis (RRMS), we performed serial optical coherence tomography (OCT) scans among a cohort of patients with RRMS on rituximab and compared rates of ganglion cell + inner plexiform layer (GCIPL) atrophy to those observed among age- and sex-matched glatiramer acetate (GA)-and natalizumab-treated patients with RRMS and healthy controls (HCs).

METHODS

In this observational study, patients with RRMS treated with a single disease-modifying therapy and HCs were followed with serial OCT for a median duration of 2.8 years. Participants with uncontrolled hypertension, diabetes mellitus, or glaucoma, and eyes with optic neuritis ≤6 months prior to baseline OCT, or during follow-up, were excluded. Statistical analyses were performed using linear mixed-effects regression.

RESULTS

During the overall follow-up period, rates of GCIPL atrophy were -0.28 ± 0.11 µm/y among rituximab-treated patients with RRMS (n = 35). This was similar to GA-treated (n = 49; -0.33 ± 0.05 µm/y; p = 0.69) and natalizumab-treated patients (n = 88; -0.17 ± 0.10 µm/y; p = 0.13) and faster than HCs (n = 78; -0.15 ± 0.03 µm/y; p = 0.006). Rituximab-treated patients exhibited 0.55 ± 0.23 µm/y faster rates of GCIPL atrophy during the first 12 months of treatment, relative to afterwards (n = 25; p = 0.02), during which period GCIPL atrophy rates were -0.14 ± 0.13 µm/y.

CONCLUSIONS

Retinal atrophy in RRMS is modulated by rituximab. Greater attenuation of retinal atrophy may occur after 12 months of rituximab treatment, following which time GCIPL atrophy rates are similar to those observed among natalizumab-treated patients with RRMS and HCs. Our findings raise the possibility that the neuroprotective therapeutic response with rituximab in RRMS may take up to 12 months, which should be confirmed by larger studies.

CLASSIFICATION OF EVIDENCE

This study provides Class IV evidence on the difference in rate of change of the GCIPL thickness in patients with RRMS comparing rituximab to other disease-modifying therapies.