Sieun Lee

Comparative analysis of repeatability of manual and automated choroidal thickness measurements in nonneovascular age-related macular degeneration.

PURPOSE We compared the reproducibility and mutual agreement of the subfoveal choroidal thickness measurements by expert raters and an automated algorithm in enhanced depth imaging optical coherence tomography (EDI-OCT) images of eyes with nonneovascular age-related macular degeneration (AMD). METHODS We recruited 44 patients with nonneovascular AMD and EDI-OCT images were acquired. Subfoveal choroidal thickness was measured manually by two expert raters and automatically by a graph-cut-based algorithm. Drusen area was measured using the automated software (version 6) of Cirrus SD-OCT. The manual and automated choroidal thickness measurements were compared in reproducibility, mutual agreement, and correlation with drusen area. RESULTS The mean subfoveal choroidal thickness was 246 ± 63 μm for the first rater, 214 ± 68 for the second rater, and 209 ± 53 for the automated algorithm. Intraclass correlation coefficients (ICC) and 95% confidence intervals (CI) were 0.96 (CI 0.94-0.98) between the raters, 0.85 (CI 0.77-0.90) between the first rater and the automated algorithm, and 0.84 (CI 0.75-0.89) between the second rater and the automated algorithm. Repeat scan measurement ICCs were 0.91 (CI 0.86-0.94) for the first rater, 0.96 (CI 0.94-0.97) for the second rater, and 0.87 (CI 0.80-0.92) for the automated algorithm. Manual and automated measurements were correlated with drusen area. CONCLUSIONS The automated algorithm generally yielded smaller choroidal thickness than the raters with a moderate level of agreement. However, its repeat scan measurement repeatability was comparable to that of the manual measurements. The mean difference between the raters indicated possible biases in different raters and rating sessions. The correlation of the automated measurements with the drusen area was comparable to that of the manual measurements. Automated subfoveal choroidal thickness measurement has potential use in clinical practice and clinical trials, with possibility for reduced time and labor cost.

Retinal angiography with real-time speckle variance optical coherence tomography

This report describes a novel, non-invasive and label-free optical imaging technique, speckle variance optical coherence tomography (svOCT), for visualising blood flow within human retinal capillary networks. This imaging system uses a custom-built swept source OCT system operating at a line rate of 100 kHz. Real-time processing and visualisation is implemented on a consumer grade graphics processing unit. To investigate the quality of microvascular detail acquired with this device we compared images of human capillary networks acquired with svOCT and fluorescein angiography. We found that the density of capillary microvasculature acquired with this svOCT device was visibly greater than fluorescein angiography. We also found that this svOCT device had the capacity to generate en face images of distinct capillary networks that are morphologically comparable with previously published histological studies. Finally, we found that this svOCT device has the ability to non-invasively illustrate the common manifestations of diabetic retinopathy and retinal vascular occlusion. The results of this study suggest that graphics processing unit accelerated svOCT has the potential to non-invasively provide useful quantitative information about human retinal capillary networks. Therefore svOCT may have clinical and research applications for the management of retinal microvascular diseases, which are a major cause of visual morbidity worldwide.

Optic Nerve Head and Peripapillary Morphometrics in Myopic Glaucoma

PURPOSE To investigate morphological characteristics of optic nerve head and peripapillary region with myopia and glaucoma. METHODS Ten normal and 17 glaucomatous myopic participants were imaged with a custom 1060-nm swept-source optical coherence tomography system. The three-dimensional images were processed and segmented for inner limiting membrane (ILM), posterior border of retinal nerve fiber layer (RNFL), Bruchs membrane (BM), and posterior border of choroid. Seven shape parameters were measured: nerve fiber layer (NFL) thickness; Bruchs membrane opening (BMO) area, eccentricity, and planarity; BMO and BM depths; and choroidal thickness. The results were analyzed by group and regional sector, and multiple regression was performed on each shape parameter with age, axial length, and glaucoma severity, measured by mean deviation (MD). RESULTS Bruchs membrane opening area (P < 0.001), eccentricity (P = 0.025), and planarity (P = 0.019) were correlated with axial length but not with MD, such that larger, more elliptical, and less planar BMO was associated with longer axial length. Several BMOs displayed a saddle-like shape configuration whose orientation appeared to be aligned with that of the BMO ellipse. All BM showed posterior deformation toward BMO such that BM closer to BMO was more posterior than that farther from BMO. Bruchs membrane depth was correlated with axial length (P = 0.014) and MD (P = 0.040) in intersubject regression, and BMO depth (P = 0.003) and BM depth (P = 0.006) were correlated with MD in intereye regression. Bruchs membrane depth was also associated with age. Choroidal thickness was negatively correlated with age (P = 0.001) and with axial length to a smaller degree (P = 0.034), but not with glaucoma severity. CONCLUSIONS Axial length was a significant factor in BMO and BM shape in normal and glaucomatous myopic subjects. Posterior deformation of BM was observed in all eyes and significantly associated with functional glaucomatous damage and age.

Longitudinal study of retinal degeneration in a rat using spectral domain optical coherence tomography.

Rodent models of retinal degenerative diseases are used by vision scientists to develop therapies and to understand mechanisms of disease progression. Measurement of changes to the thickness of the various retinal layers provides an objective metric to evaluate the performance of the therapy. Because invasive histology is terminal and provides only a single data point, non-invasive imaging modalities are required to better study progression, and to reduce the number of animals used in research. Optical Coherence Tomography (OCT) has emerged as a dominant imaging modality for human ophthalmic imaging, but has only recently gained significant attention for rodent retinal imaging. OCT provides cross section images of retina with micron-scale resolution which permits measurement of the retinal layer thickness. However, in order to be useful to vision scientists, a significant fraction of the retinal surface needs to be measured. In addition, because the retinal thickness normally varies as a function of distance from optic nerve head, it is critical to sample all regions of the retina in a systematic fashion. We present a longitudinal study of OCT to measure retinal degeneration in rats which have undergone optic nerve axotomy, a well characterized form of rapid retinal degeneration. Volumetric images of the retina acquired with OCT in a time course study were segmented in 2D using a semi-automatic segmentation algorithm. Then, using a 3D algorithm, thickness measurements were quantified across the surface of the retina for all volume segmentations. The resulting maps of the changes to retinal thickness over time represent the progression of degeneration across the surface of the retina during injury. The computational tools complement OCT retinal volumetric acquisition, resulting in a powerful tool for vision scientists working with rodents.

Segmentation of the foveal microvasculature using deep learning networks

Abstract. Accurate segmentation of the retinal microvasculature is a critical step in the quantitative analysis of the retinal circulation, which can be an important marker in evaluating the severity of retinal diseases. As manual segmentation remains the gold standard for segmentation of optical coherence tomography angiography (OCT-A) images, we present a method for automating the segmentation of OCT-A images using deep neural networks (DNNs). Eighty OCT-A images of the foveal region in 12 eyes from 6 healthy volunteers were acquired using a prototype OCT-A system and subsequently manually segmented. The automated segmentation of the blood vessels in the OCT-A images was then performed by classifying each pixel into vessel or nonvessel class using deep convolutional neural networks. When the automated results were compared against the manual segmentation results, a maximum mean accuracy of 0.83 was obtained. When the automated results were compared with inter and intrarater accuracies, the automated results were shown to be comparable to the human raters suggesting that segmentation using DNNs is comparable to a second manual rater. As manually segmenting the retinal microvasculature is a tedious task, having a reliable automated output such as automated segmentation by DNNs, is an important step in creating an automated output.

Automatic detection of subretinal fluid and sub-retinal pigment epithelium fluid in optical coherence tomography images

Age-related macular degeneration (AMD) is the leading cause of blindness in developed countries. Subretinal fluid (SRF) and sub-retinal pigment epithelium (sub-RPE) fluid are signs of AMD and can be detected in optical coherence tomography images. However, manual detection and segmentation of SRFs and sub-RPE fluids are laborious and time consuming. In this paper, a novel pipeline is proposed for automatic detection of SRFs and sub-RPE fluids. First, top and bottom layers of retina are segmented using a graph cut method. Then, a Split Bregman-based segmentation method is used to segment dark regions between layers. These segmented regions are considered as potential fluid candidates, on which a set of features are generated. After that, a random forest classifier is trained to distinguish between the true fluid regions from the falsely detected fluid regions. This method shows reasonable performance in a leave-one-out evaluation using a dataset from 21 patients.

Exact Surface Registration of Retinal Surfaces From 3-D Optical Coherence Tomography Images

Nonrigid registration of optical coherence tomography (OCT) images is an important problem in studying eye diseases, evaluating the effect of pharmaceuticals in treating vision loss, and performing group-wise cross-sectional analysis. High dimensional nonrigid registration algorithms required for cross-sectional and longitudinal analysis are still being developed for accurate registration of OCT image volumes, with the speckle noise in images presenting a challenge for registration. Development of algorithms for segmentation of OCT images to generate surface models of retinal layers has advanced considerably and several algorithms are now available that can segment retinal OCT images into constituent retinal surfaces. Important morphometric measurements can be extracted if accurate surface registration algorithm for registering retinal surfaces onto corresponding template surfaces were available. In this paper, we present a novel method to perform multiple and simultaneous retinal surface registration, targeted to registering surfaces extracted from ocular volumetric OCT images. This enables a point-to-point correspondence (homology) between template and subject surfaces, allowing for a direct, vertex-wise comparison of morphometric measurements across subject groups. We demonstrate that this approach can be used to localize and analyze regional changes in choroidal and nerve fiber layer thickness among healthy and glaucomatous subjects, allowing for cross-sectional population wise analysis. We also demonstrate the methods ability to track longitudinal changes in optic nerve head morphometry, allowing for within-individual tracking of morphometric changes. This method can also, in the future, be used as a precursor to 3-D OCT image registration to better initialize nonrigid image registration algorithms closer to the desired solution.

Atlas-based shape analysis and classification of retinal optical coherence tomography images using the functional shape (fshape) framework

&NA; We propose a novel approach for quantitative shape variability analysis in retinal optical coherence tomography images using the functional shape (fshape) framework. The fshape framework uses surface geometry together with functional measures, such as retinal layer thickness defined on the layer surface, for registration across anatomical shapes. This is used to generate a population mean template of the geometry‐function measures from each individual. Shape variability across multiple retinas can be measured by the geometrical deformation and functional residual between the template and each of the observations. To demonstrate the clinical relevance and application of the framework, we generated atlases of the inner layer surface and layer thickness of the Retinal Nerve Fiber Layer (RNFL) of glaucomatous and normal subjects, visualizing detailed spatial pattern of RNFL loss in glaucoma. Additionally, a regularized linear discriminant analysis classifier was used to automatically classify glaucoma, glaucoma‐suspect, and control cases based on RNFL fshape metrics.

Comparison of the clinical disc margin seen in stereo disc photographs with neural canal opening seen in optical coherence tomography images.

Purpose:To compare neural canal opening (NCO) with the clinical optic disc margin (DM) seen and to investigate the planarity of the NCO in normal human optic nerve heads (ONH). Methods:Sixteen eyes were imaged. Twelve healthy eyes were selected for planarity and 9 for NCO and DM correspondence. All subjects were subjected to a visual field examination, stereo disc photograph (SDP), scanning laser ophthalmoscopy, clinical examination with a fellowship trained glaucoma specialist, and optical coherence tomography imaging. Three reviewers delineated the NCO and inner limiting membrane on optical coherence tomography images. The clinical DM was delineated by a glaucoma specialist while viewing SDPs. Plane error was calculated for NCO and for Bruch membrane (BM) at distances 80 and 120 &mgr;m from NCO. Results:The NCO segmentation interrater variability was low with an average coefficient of variation of 2.7%. A regional variation of the SDP and NCO correspondence was observed, wherein the temporal region had the largest coefficient of variation. The plane error of the NCO and BM were similar and was approximately 12 &mgr;m, which is small relative to an average DM diameter of 1.7 mm. Conclusions:The BM opening has a good correspondence with the clinical DM seen in SDPs. NCO delineation seemed to be reliable. The BM and NCO are relatively planar in normal humans and can be further evaluated for longitudinal studies to observe stability.

Rapid radial optical coherence tomography image acquisition

Abstract. We demonstrate how compressive sampling can be used to expedite volumetric optical coherence tomography (OCT) image acquisition. We propose a novel method to interpolate OCT volumetric images from data acquired by radial B-scans in the Cartesian coordinate system. Due to the inherent polar symmetry in the human eye, the (r, θ, z) coordinate system provides a natural domain to perform the interpolation. We demonstrate that the method has minimal effect on image quality even when up to 88% of the data is not acquired. The potential outcome of this work could lead to significant reductions in OCT volume acquisition time in clinical practice.