Active contour method for ILM segmentation in ONH volume scans in retinal OCT

[MiHUBx: a digital progress hub for the use of intersectoral clinical data sets using the example of diabetic macular edema]

BACKGROUND

Evidence-based treatment recommendations are helpful in the corresponding discipline-specific treatment but can hardly take data from real-world care into account. In order to make better use of this in everyday clinical practice, including with respect to predictive statements about disease development or treatment success, models with data from treatment must be developed in order to use them for the development of assistive artificial intelligence.

GOAL

The aim of the Use Case 1 of the medical informatics hub in Saxony (MiHUBx) is the development of a model based on treatment and research data for a treatment algorithm supported by biomarkers and also the development of the necessary digital infrastructure.

MATERIAL AND METHODS

Step by step, the necessary partners in hospitals and practices will be brought together technically or through research questions within Use Case 1 "Ophthalmology meets Diabetology", a regional digital progress hub in health, the medical informatics hub in Saxony (MiHUBx ) of the nationwide medical informatics initiative (MII).

RESULTS

Based on joint studies with diabetologists, robust serological and imaging biomarkers were selected that provide evidence of the development of diabetic macular edema (DME). In the future, these and other scientifically proven prognostic markers will be incorporated into a treatment algorithm that is supported by artificial intelligence (AI). For this purpose, model procedures are being developed together with medical informatics specialists. At the same time, a data integration center (DIZ) was established.

CONCLUSION

In addition to the structured and technical combination of the previously disseminated and partially heterogeneous treatment data, the Use Case 1 defines the chances and hurdles for using such real-world data to develop artificial intelligence.

Unraveling the complexity of Optical Coherence Tomography image segmentation using machine and deep learning techniques: A review

Optical Coherence Tomography (OCT) is an emerging technology that provides three-dimensional images of the microanatomy of biological tissue in-vivo and at micrometer-scale resolution. OCT imaging has been widely used to diagnose and manage various medical diseases, such as macular degeneration, glaucoma, and coronary artery disease. Despite its wide range of applications, the segmentation of OCT images remains difficult due to the complexity of tissue structures and the presence of artifacts. In recent years, different approaches have been used for OCT image segmentation, such as intensity-based, region-based, and deep learning-based methods. This paper reviews the major advances in state-of-the-art OCT image segmentation techniques. It provides an overview of the advantages and limitations of each method and presents the most relevant research works related to OCT image segmentation. It also provides an overview of existing datasets and discusses potential clinical applications. Additionally, this review gives an in-depth analysis of machine learning and deep learning approaches for OCT image segmentation. It outlines challenges and opportunities for further research in this field.

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.

Annotation-efficient learning for OCT segmentation

Deep learning has been successfully applied to OCT segmentation. However, for data from different manufacturers and imaging protocols, and for different regions of interest (ROIs), it requires laborious and time-consuming data annotation and training, which is undesirable in many scenarios, such as surgical navigation and multi-center clinical trials. Here we propose an annotation-efficient learning method for OCT segmentation that could significantly reduce annotation costs. Leveraging self-supervised generative learning, we train a Transformer-based model to learn the OCT imagery. Then we connect the trained Transformer-based encoder to a CNN-based decoder, to learn the dense pixel-wise prediction in OCT segmentation. These training phases use open-access data and thus incur no annotation costs, and the pre-trained model can be adapted to different data and ROIs without re-training. Based on the greedy approximation for the k-center problem, we also introduce an algorithm for the selective annotation of the target data. We verified our method on publicly-available and private OCT datasets. Compared to the widely-used U-Net model with 100% training data, our method only requires ∼10% of the data for achieving the same segmentation accuracy, and it speeds the training up to ∼3.5 times. Furthermore, our proposed method outperforms other potential strategies that could improve annotation efficiency. We think this emphasis on learning efficiency may help improve the intelligence and application penetration of OCT-based technologies.

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.

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.

Automatic Identification and Intuitive Map Representation of the Epiretinal Membrane Presence in 3D OCT Volumes

Optical Coherence Tomography (OCT) is a medical image modality providing high-resolution cross-sectional visualizations of the retinal tissues without any invasive procedure, commonly used in the analysis of retinal diseases such as diabetic retinopathy or retinal detachment. Early identification of the epiretinal membrane (ERM) facilitates ERM surgical removal operations. Moreover, presence of the ERM is linked to other retinal pathologies, such as macular edemas, being among the main causes of vision loss. In this work, we propose an automatic method for the characterization and visualization of the ERM's presence using 3D OCT volumes. A set of 452 features is refined using the Spatial Uniform ReliefF (SURF) selection strategy to identify the most relevant ones. Afterwards, a set of representative classifiers is trained, selecting the most proficient model, generating a 2D reconstruction of the ERM's presence. Finally, a post-processing stage using a set of morphological operators is performed to improve the quality of the generated maps. To verify the proposed methodology, we used 20 3D OCT volumes, both with and without the ERM's presence, totalling 2428 OCT images manually labeled by a specialist. The most optimal classifier in the training stage achieved a mean accuracy of 91 . 9 % . Regarding the post-processing stage, mean specificity values of 91 . 9 % and 99 . 0 % were obtained from volumes with and without the ERM's presence, respectively.

Automatic identification and characterization of the epiretinal membrane in OCT images

Optical coherence tomography (OCT) is a medical image modality that is used to capture, non-invasively, high-resolution cross-sectional images of the retinal tissue. These images constitute a suitable scenario for the diagnosis of relevant eye diseases like the vitreomacular traction or the diabetic retinopathy. The identification of the epiretinal membrane (ERM) is a relevant issue as its presence constitutes a symptom of diseases like the macular edema, deteriorating the vision quality of the patients. This work presents an automatic methodology for the identification of the ERM presence in OCT scans. Initially, a complete and heterogeneous set of features was defined to capture the properties of the ERM in the OCT scans. Selected features went through a feature selection process to further improve the method efficiency. Additionally, representative classifiers were trained and tested to measure the suitability of the proposed approach. The method was tested with a dataset of 285 OCT scans labeled by a specialist. In particular, 3,600 samples were equally extracted from the dataset, representing zones with and without ERM presence. Different experiments were conducted to reach the most suitable approach. Finally, selected classifiers were trained and compared using different metrics, providing in the best configuration an accuracy of 89.35%.

Optic nerve head three-dimensional shape analysis

We present a method for optic nerve head (ONH) 3-D shape analysis from retinal optical coherence tomography (OCT). The possibility to noninvasively acquire in vivo high-resolution 3-D volumes of the ONH using spectral domain OCT drives the need to develop tools that quantify the shape of this structure and extract information for clinical applications. The presented method automatically generates a 3-D ONH model and then allows the computation of several 3-D parameters describing the ONH. The method starts with a high-resolution OCT volume scan as input. From this scan, the model-defining inner limiting membrane (ILM) as inner surface and the retinal pigment epithelium as outer surface are segmented, and the Bruch's membrane opening (BMO) as the model origin is detected. Based on the generated ONH model by triangulated 3-D surface reconstruction, different parameters (areas, volumes, annular surface ring, minimum distances) of different ONH regions can then be computed. Additionally, the bending energy (roughness) in the BMO region on the ILM surface and 3-D BMO-MRW surface area are computed. We show that our method is reliable and robust across a large variety of ONH topologies (specific to this structure) and present a first clinical application.