Retinal layer parcellation of optical coherence tomography images: Data resource for multiple sclerosis and healthy controls

HDB-Net: hierarchical dual-branch network for retinal layer segmentation in diseased OCT images

Optical coherence tomography (OCT) retinal layer segmentation is a critical procedure of the modern ophthalmic process, which can be used for diagnosis and treatment of diseases such as diabetic macular edema (DME) and multiple sclerosis (MS). Due to the difficulties of low OCT image quality, highly similar retinal interlayer morphology, and the uncertain presence, shape and size of lesions, the existing algorithms do not perform well. In this work, we design an HDB-Net network for retinal layer segmentation in diseased OCT images, which solves this problem by combining global and detailed features. First, the proposed network uses a Swin transformer and Res50 as a parallel backbone network, combined with the pyramid structure in UperNet, to extract global context and aggregate multi-scale information from images. Secondly, a feature aggregation module (FAM) is designed to extract global context information from the Swin transformer and local feature information from ResNet by introducing mixed attention mechanism. Finally, the boundary awareness and feature enhancement module (BA-FEM) is used to extract the retinal layer boundary information and topological order from the low-resolution features of the shallow layer. Our approach has been validated on two public datasets, and Dice scores were 87.61% and 92.44, respectively, both outperforming other state-of-the-art technologies.

Discrimination of multiple sclerosis using scanning laser ophthalmoscopy images with autoencoder-based feature extraction

OBJECTIVE

Optical coherence tomography (OCT) investigations have revealed that the thickness of inner retinal layers becomes decreased in multiple sclerosis (MS) patients, compared to healthy control (HC) individuals. To date, a number of studies have applied machine learning to OCT thickness measurements, aiming to enable accurate and automated diagnosis of the disease. However, there have much less emphasis on other less common retinal imaging modalities, like infrared scanning laser ophthalmoscopy (IR-SLO), for classifying MS. IR-SLO uses laser light to capture high-resolution fundus images, often performed in conjunction with OCT to lock B-scans at a fixed position.

METHODS

We incorporated two independent datasets of IR-SLO images from the Isfahan and Johns Hopkins centers, consisting of 164 MS and 150 HC images. A subject-wise data splitting approach was employed to ensure that there was no leakage between training and test datasets. Several state-of-the-art convolutional neural networks (CNNs), including VGG-16, VGG-19, ResNet-50, and InceptionV3, and a CNN with a custom architecture were employed. In the next step, we designed a convolutional autoencoder (CAE) to extract semantic features subsequently given as inputs to four conventional ML classifiers, including support vector machine (SVM), k-nearest neighbor (K-NN), random forest (RF), and multi-layer perceptron (MLP).

RESULTS

The custom CNN (85 % accuracy, 85 % sensitivity, 87 % specificity, 93 % area under the receiver operating characteristics [AUROC], and 94 % area under the precision-recall curve [AUPRC]) outperformed state-of-the-art models (84 % accuracy, 83 % sensitivity, 87 % specificity, 92 % AUROC, and 94 % AUPRC); however, utilizing a combination of the CAE and MLP yields even superior results (88 % accuracy, 86 % sensitivity, 91 % specificity, 94 % AUROC, and 95 % AUPRC).

CONCLUSIONS

We utilized IR-SLO images to differentiate between MS and HC eyes, with promising results achieved using a combination of CAE and MLP. Future multi-center studies involving more heterogenous data are necessary to assess the feasibility of integrating IR-SLO images into routine clinical practice.

Exploring Publicly Accessible Optical Coherence Tomography Datasets: A Comprehensive Overview

Artificial intelligence has transformed medical diagnostic capabilities, particularly through medical image analysis. AI algorithms perform well in detecting abnormalities with a strong performance, enabling computer-aided diagnosis by analyzing the extensive amounts of patient data. The data serve as a foundation upon which algorithms learn and make predictions. Thus, the importance of data cannot be underestimated, and clinically corresponding datasets are required. Many researchers face a lack of medical data due to limited access, privacy concerns, or the absence of available annotations. One of the most widely used diagnostic tools in ophthalmology is Optical Coherence Tomography (OCT). Addressing the data availability issue is crucial for enhancing AI applications in the field of OCT diagnostics. This review aims to provide a comprehensive analysis of all publicly accessible retinal OCT datasets. Our main objective is to compile a list of OCT datasets and their properties, which can serve as an accessible reference, facilitating data curation for medical image analysis tasks. For this review, we searched through the Zenodo repository, Mendeley Data repository, MEDLINE database, and Google Dataset search engine. We systematically evaluated all the identified datasets and found 23 open-access datasets containing OCT images, which significantly vary in terms of size, scope, and ground-truth labels. Our findings indicate the need for improvement in data-sharing practices and standardized documentation. Enhancing the availability and quality of OCT datasets will support the development of AI algorithms and ultimately improve diagnostic capabilities in ophthalmology. By providing a comprehensive list of accessible OCT datasets, this review aims to facilitate better utilization and development of AI in medical image analysis.

Hybrid deep learning and optimal graph search method for optical coherence tomography layer segmentation in diseases affecting the optic nerve

Accurate segmentation of retinal layers in optical coherence tomography (OCT) images is critical for assessing diseases that affect the optic nerve, but existing automated algorithms often fail when pathology causes irregular layer topology, such as extreme thinning of the ganglion cell-inner plexiform layer (GCIPL). Deep LOGISMOS, a hybrid approach that combines the strengths of deep learning and 3D graph search to overcome their limitations, was developed to improve the accuracy, robustness and generalizability of retinal layer segmentation. The method was trained on 124 OCT volumes from both eyes of 31 non-arteritic anterior ischemic optic neuropathy (NAION) patients and tested on three cross-sectional datasets with available reference tracings: Test-NAION (40 volumes from both eyes of 20 NAION subjects), Test-G (29 volumes from 29 glaucoma subjects/eyes), and Test-JHU (35 volumes from 21 multiple sclerosis and 14 control subjects/eyes) and one longitudinal dataset without reference tracings: Test-G-L (155 volumes from 15 glaucoma patients/eyes). In the three test datasets with reference tracings (Test-NAION, Test-G, and Test-JHU), Deep LOGISMOS achieved very high Dice similarity coefficients (%) on GCIPL: 89.97±3.59, 90.63±2.56, and 94.06±1.76, respectively. In the same context, Deep LOGISMOS outperformed the Iowa reference algorithms by improving the Dice score by 17.5, 5.4, and 7.5, and also surpassed the deep learning framework nnU-Net with improvements of 4.4, 3.7, and 1.0. For the 15 severe glaucoma eyes with marked GCIPL thinning (Test-G-L), it demonstrated reliable regional GCIPL thickness measurement over five years. The proposed Deep LOGISMOS approach has potential to enhance precise quantification of retinal structures, aiding diagnosis and treatment management of optic nerve diseases.

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.

OCTDL: Optical Coherence Tomography Dataset for Image-Based Deep Learning Methods

Optical coherence tomography (OCT) is a non-invasive imaging technique with extensive clinical applications in ophthalmology. OCT enables the visualization of the retinal layers, playing a vital role in the early detection and monitoring of retinal diseases. OCT uses the principle of light wave interference to create detailed images of the retinal microstructures, making it a valuable tool for diagnosing ocular conditions. This work presents an open-access OCT dataset (OCTDL) comprising over 2000 OCT images labeled according to disease group and retinal pathology. The dataset consists of OCT records of patients with Age-related Macular Degeneration (AMD), Diabetic Macular Edema (DME), Epiretinal Membrane (ERM), Retinal Artery Occlusion (RAO), Retinal Vein Occlusion (RVO), and Vitreomacular Interface Disease (VID). The images were acquired with an Optovue Avanti RTVue XR using raster scanning protocols with dynamic scan length and image resolution. Each retinal b-scan was acquired by centering on the fovea and interpreted and cataloged by an experienced retinal specialist. In this work, we applied Deep Learning classification techniques to this new open-access dataset.

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.

Reproducibility evaluation of the effects of MRI defacing on brain segmentation

Purpose

Recent advances in magnetic resonance (MR) scanner quality and the rapidly improving nature of facial recognition software have necessitated the introduction of MR defacing algorithms to protect patient privacy. As a result, there are a number of MR defacing algorithms available to the neuroimaging community, with several appearing in just the last 5 years. While some qualities of these defacing algorithms, such as patient identifiability, have been explored in the previous works, the potential impact of defacing on neuroimage processing has yet to be explored.

Approach

We qualitatively evaluate eight MR defacing algorithms on 179 subjects from the OASIS-3 cohort and 21 subjects from the Kirby-21 dataset. We also evaluate the effects of defacing on two neuroimaging pipelines-SLANT and FreeSurfer-by comparing the segmentation consistency between the original and defaced images.

Results

Defacing can alter brain segmentation and even lead to catastrophic failures, which are more frequent with some algorithms, such as Quickshear, MRI_Deface, and FSL_deface. Compared to FreeSurfer, SLANT is less affected by defacing. On outputs that pass the quality check, the effects of defacing are less pronounced than those of rescanning, as measured by the Dice similarity coefficient.

Conclusions

The effects of defacing are noticeable and should not be disregarded. Extra attention, in particular, should be paid to the possibility of catastrophic failures. It is crucial to adopt a robust defacing algorithm and perform a thorough quality check before releasing defaced datasets. To improve the reliability of analysis in scenarios involving defaced MRIs, it is encouraged to include multiple brain segmentation pipelines.

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.

Exploiting multi-granularity visual features for retinal layer segmentation in human eyes

Accurate segmentation of retinal layer boundaries can facilitate the detection of patients with early ophthalmic disease. Typical segmentation algorithms operate at low resolutions without fully exploiting multi-granularity visual features. Moreover, several related studies do not release their datasets that are key for the research on deep learning-based solutions. We propose a novel end-to-end retinal layer segmentation network based on ConvNeXt, which can retain more feature map details by using a new depth-efficient attention module and multi-scale structures. In addition, we provide a semantic segmentation dataset containing 206 retinal images of healthy human eyes (named NR206 dataset), which is easy to use as it does not require any additional transcoding processing. We experimentally show that our segmentation approach outperforms state-of-the-art approaches on this new dataset, achieving, on average, a Dice score of 91.3% and mIoU of 84.4%. Moreover, our approach achieves state-of-the-art performance on a glaucoma dataset and a diabetic macular edema (DME) dataset, showing that our model is also suitable for other applications. We will make our source code and the NR206 dataset publicly available at (https://github.com/Medical-Image-Analysis/Retinal-layer-segmentation).

A reproducibility evaluation of the effects of MRI defacing on brain segmentation

Purpose

Recent advances in magnetic resonance (MR) scanner quality and the rapidly improving nature of facial recognition software have necessitated the introduction of MR defacing algorithms to protect patient privacy. As a result, there are a number of MR defacing algorithms available to the neuroimaging community, with several appearing in just the last five years. While some qualities of these defacing algorithms, such as patient identifiability, have been explored in previous works, the potential impact of defacing on neuroimage processing has yet to be explored.

Approach

We qualitatively evaluate eight MR defacing algorithms on 179 subjects from the OASIS-3 cohort and the 21 subjects from the Kirby-21 dataset. We also evaluate the effects of defacing on two neuroimaging pipelines-SLANT and FreeSurfer-by comparing the segmentation consistency between the original and defaced images.

Results

Defacing can alter brain segmentation and even lead to catastrophic failures, which are more frequent with some algorithms such as Quickshear, MRI_Deface, and FSL_deface. Compared to FreeSurfer, SLANT is less affected by defacing. On outputs that pass the quality check, the effects of defacing are less pronounced than those of rescanning, as measured by the Dice similarity coefficient.

Conclusions

The effects of defacing are noticeable and should not be disregarded. Extra attention, in particular, should be paid to the possibility of catastrophic failures. It is crucial to adopt a robust defacing algorithm and perform a thorough quality check before releasing defaced datasets. To improve the reliability of analysis in scenarios involving defaced MRIs, it's encouraged to include multiple brain segmentation pipelines.

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.

Boundary-enhanced semi-supervised retinal layer segmentation in optical coherence tomography images using fewer labels

Automatic segmentation of multiple layers in retinal optical coherence tomography (OCT) images is crucial for eye disease diagnosis and treatment. Despite the success of deep learning algorithms, it still remains a challenge due to the blurry layer boundaries and lack of adequate pixel-wise annotations. To tackle these issues, we propose a Boundary-Enhanced Semi-supervised Network (BE-SemiNet) that exploits an auxiliary distance map regression task to improve retinal layer segmentation with scarce labeled data and abundant unlabeled data. Specifically, a novel Unilaterally Truncated Distance Map (UTDM) is firstly introduced to alleviate the class imbalance problem and enhance the layer boundary learning in the regression task. Then for the pixel-wise segmentation and UTDM regression branches, we impose task-level and data-level consistency regularization on unlabeled data to enrich the diversity of unsupervised information and improve the regularization effects. Pseudo supervision is incorporated in consistency regularization to bridge the task prediction spaces for consistency and expand training labeled data. Experiments on two public retinal OCT datasets show that our method can greatly improve the supervised baseline performance with only 5 annotations and outperform the state-of-the-art methods. Since it is difficult and labor-expensive to obtain adequate pixel-wise annotations in practice, our method has a promising application future in clinical retinal OCT image analysis.

Multiscale joint segmentation method for retinal optical coherence tomography images using a bidirectional wave algorithm and improved graph theory

Morphology and functional metrics of retinal layers are important biomarkers for many human ophthalmic diseases. Automatic and accurate segmentation of retinal layers is crucial for disease diagnosis and research. To improve the performance of retinal layer segmentation, a multiscale joint segmentation framework for retinal optical coherence tomography (OCT) images based on bidirectional wave algorithm and improved graph theory is proposed. In this framework, the bidirectional wave algorithm was used to segment edge information in multiscale images, and the improved graph theory was used to modify edge information globally, to realize automatic and accurate segmentation of eight retinal layer boundaries. This framework was tested on two public datasets and two OCT imaging systems. The test results show that, compared with other state-of-the-art methods, this framework does not need data pre-training and parameter pre-adjustment on different datasets, and can achieve sub-pixel retinal layer segmentation on a low-configuration computer.

Wavelet attention network for the segmentation of layer structures on OCT images

Automatic segmentation of layered tissue is critical for optical coherence tomography (OCT) image analysis. The development of deep learning techniques provides various solutions to this problem, while most existing methods suffer from topological errors such as outlier prediction and label disconnection. The channel attention mechanism is a powerful technique to address these problems due to its simplicity and robustness. However, it relies on global average pooling (GAP), which only calculates the lowest frequency component and leaves other potentially useful information unexplored. In this study, we use the discrete wavelet transform (DWT) to extract multi-spectral information and propose the wavelet attention network (WATNet) for tissue layer segmentation. The DWT-based attention mechanism enables multi-spectral analysis with no complex frequency-selection process and can be easily embedded to existing frameworks. Furthermore, the various wavelet bases make the WATNet adaptable to different tasks. Experiments on a self-collected esophageal dataset and two public retinal OCT dataset demonstrated that the WATNet achieved better performance compared to several widely used deep networks, confirming the advantages of the proposed method.

DLTTA: Dynamic Learning Rate for Test-Time Adaptation on Cross-Domain Medical Images

Test-time adaptation (TTA) has increasingly been an important topic to efficiently tackle the cross-domain distribution shift at test time for medical images from different institutions. Previous TTA methods have a common limitation of using a fixed learning rate for all the test samples. Such a practice would be sub-optimal for TTA, because test data may arrive sequentially therefore the scale of distribution shift would change frequently. To address this problem, we propose a novel dynamic learning rate adjustment method for test-time adaptation, called DLTTA, which dynamically modulates the amount of weights update for each test image to account for the differences in their distribution shift. Specifically, our DLTTA is equipped with a memory bank based estimation scheme to effectively measure the discrepancy of a given test sample. Based on this estimated discrepancy, a dynamic learning rate adjustment strategy is then developed to achieve a suitable degree of adaptation for each test sample. The effectiveness and general applicability of our DLTTA is extensively demonstrated on three tasks including retinal optical coherence tomography (OCT) segmentation, histopathological image classification, and prostate 3D MRI segmentation. Our method achieves effective and fast test-time adaptation with consistent performance improvement over current state-of-the-art test-time adaptation methods. Code is available at https://github.com/med-air/DLTTA.

LOCTseg: A lightweight fully convolutional network for end-to-end optical coherence tomography segmentation

This article presents a novel end-to-end automatic solution for semantic segmentation of optical coherence tomography (OCT) images. OCT is a non-invasive imaging technology widely used in clinical practice due to its ability to acquire high-resolution cross-sectional images of the ocular fundus. Due to the large variability of the retinal structures, OCT segmentation is usually carried out manually and requires expert knowledge. This study introduces a novel fully convolutional network (FCN) architecture designated by LOCTSeg, for end-to-end automatic segmentation of diagnostic markers in OCT b-scans. LOCTSeg is a lightweight deep FCN optimized for balancing performance and efficiency. Unlike state-of-the-art FCNs used in image segmentation, LOCTSeg achieves competitive inference speed without sacrificing segmentation accuracy. The proposed LOCTSeg is evaluated on two publicly available benchmarking datasets: (1) annotated retinal OCT image database (AROI) comprising 1136 images, and (2) healthy controls and multiple sclerosis lesions (HCMS) consisting of 1715 images. Moreover, we evaluated the proposed LOCTSeg with a private dataset of 250 OCT b-scans acquired from epiretinal membrane (ERM) and healthy patients. Results of the evaluation demonstrate empirically the effectiveness of the proposed algorithm, which improves the state-of-the-art Dice score from 69% to 73% and from 91% to 92% on AROI and HCMS datasets, respectively. Furthermore, LOCTSeg outperforms comparable lightweight FCNs' Dice score by margins between 4% and 15% on ERM segmentation.

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.

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.

Recent Advanced Deep Learning Architectures for Retinal Fluid Segmentation on Optical Coherence Tomography Images

With non-invasive and high-resolution properties, optical coherence tomography (OCT) has been widely used as a retinal imaging modality for the effective diagnosis of ophthalmic diseases. The retinal fluid is often segmented by medical experts as a pivotal biomarker to assist in the clinical diagnosis of age-related macular diseases, diabetic macular edema, and retinal vein occlusion. In recent years, the advanced machine learning methods, such as deep learning paradigms, have attracted more and more attention from academia in the retinal fluid segmentation applications. The automatic retinal fluid segmentation based on deep learning can improve the semantic segmentation accuracy and efficiency of macular change analysis, which has potential clinical implications for ophthalmic pathology detection. This article summarizes several different deep learning paradigms reported in the up-to-date literature for the retinal fluid segmentation in OCT images. The deep learning architectures include the backbone of convolutional neural network (CNN), fully convolutional network (FCN), U-shape network (U-Net), and the other hybrid computational methods. The article also provides a survey on the prevailing OCT image datasets used in recent retinal segmentation investigations. The future perspectives and some potential retinal segmentation directions are discussed in the concluding context.

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.

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.

Autoencoder based self-supervised test-time adaptation for medical image analysis

Deep neural networks have been successfully applied to medical image analysis tasks like segmentation and synthesis. However, even if a network is trained on a large dataset from the source domain, its performance on unseen test domains is not guaranteed. The performance drop on data obtained differently from the network's training data is a major problem (known as domain shift) in deploying deep learning in clinical practice. Existing work focuses on retraining the model with data from the test domain, or harmonizing the test domain's data to the network training data. A common practice is to distribute a carefully-trained model to multiple users (e.g., clinical centers), and then each user uses the model to process their own data, which may have a domain shift (e.g., varying imaging parameters and machines). However, the lack of availability of the source training data and the cost of training a new model often prevents the use of known methods to solve user-specific domain shifts. Here, we ask whether we can design a model that, once distributed to users, can quickly adapt itself to each new site without expensive retraining or access to the source training data? In this paper, we propose a model that can adapt based on a single test subject during inference. The model consists of three parts, which are all neural networks: a task model (T) which performs the image analysis task like segmentation; a set of autoencoders (AEs); and a set of adaptors (As). The task model and autoencoders are trained on the source dataset and can be computationally expensive. In the deployment stage, the adaptors are trained to transform the test image and its features to minimize the domain shift as measured by the autoencoders' reconstruction loss. Only the adaptors are optimized during the testing stage with a single test subject thus is computationally efficient. The method was validated on both retinal optical coherence tomography (OCT) image segmentation and magnetic resonance imaging (MRI) T1-weighted to T2-weighted image synthesis. Our method, with its short optimization time for the adaptors (10 iterations on a single test subject) and its additional required disk space for the autoencoders (around 15 MB), can achieve significant performance improvement. Our code is publicly available at: https://github.com/YufanHe/self-domain-adapted-network.

Structured layer surface segmentation for retina OCT using fully convolutional regression networks

Optical coherence tomography (OCT) is a noninvasive imaging modality with micrometer resolution which has been widely used for scanning the retina. Retinal layers are important biomarkers for many diseases. Accurate automated algorithms for segmenting smooth continuous layer surfaces with correct hierarchy (topology) are important for automated retinal thickness and surface shape analysis. State-of-the-art methods typically use a two step process. Firstly, a trained classifier is used to label each pixel into either background and layers or boundaries and non-boundaries. Secondly, the desired smooth surfaces with the correct topology are extracted by graph methods (e.g., graph cut). Data driven methods like deep networks have shown great ability for the pixel classification step, but to date have not been able to extract structured smooth continuous surfaces with topological constraints in the second step. In this paper, we combine these two steps into a unified deep learning framework by directly modeling the distribution of the surface positions. Smooth, continuous, and topologically correct surfaces are obtained in a single feed forward operation. The proposed method was evaluated on two publicly available data sets of healthy controls and subjects with either multiple sclerosis or diabetic macular edema, and is shown to achieve state-of-the art performance with sub-pixel accuracy.

Fully Convolutional Boundary Regression for Retina OCT Segmentation

A major goal of analyzing retinal optical coherence tomography (OCT) images is retinal layer segmentation. Accurate automated algorithms for segmenting smooth continuous layer surfaces, with correct hierarchy (topology) are desired for monitoring disease progression. State-of-the-art methods use a trained classifier to label each pixel into background, layer, or surface pixels. The final step of extracting the desired smooth surfaces with correct topology are mostly performed by graph methods (e.g. shortest path, graph cut). However, manually building a graph with varying constraints by retinal region and pathology and solving the minimization with specialized algorithms will degrade the flexibility and time efficiency of the whole framework. In this paper, we directly model the distribution of surface positions using a deep network with a fully differentiable soft argmax to obtain smooth, continuous surfaces in a single feed forward operation. A special topology module is used in the deep network both in the training and testing stages to guarantee the surface topology. An extra deep network output branch is also used for predicting lesion and layers in a pixel-wise labeling scheme. The proposed method was evaluated on two publicly available data sets of healthy controls, subjects with multiple sclerosis, and diabetic macular edema; it achieves state-of-the art sub-pixel results.

Deep learning based topology guaranteed surface and MME segmentation of multiple sclerosis subjects from retinal OCT

Optical coherence tomography (OCT) is a noninvasive imaging modality that can be used to obtain depth images of the retina. Patients with multiple sclerosis (MS) have thinning retinal nerve fiber and ganglion cell layers, and approximately 5% of MS patients will develop microcystic macular edema (MME) within the retina. Segmentation of both the retinal layers and MME can provide important information to help monitor MS progression. Graph-based segmentation with machine learning preprocessing is the leading method for retinal layer segmentation, providing accurate surface delineations with the correct topological ordering. However, graph methods are time-consuming and they do not optimally incorporate joint MME segmentation. This paper presents a deep network that extracts continuous, smooth, and topology-guaranteed surfaces and MMEs. The network learns shape priors automatically during training rather than being hard-coded as in graph methods. In this new approach, retinal surfaces and MMEs are segmented together with two cascaded deep networks in a single feed forward propagation. The proposed framework obtains retinal surfaces (separating the layers) with sub-pixel surface accuracy comparable to the best existing graph methods and MMEs with better accuracy than the state-of-the-art method. The full segmentation operation takes only ten seconds for a 3D volume.