Three-Dimensional Analysis of Retinal Layer Texture: Identification of Fluid-Filled Regions in SD-OCT of the Macula

Dept. of Ophthalmology & Visual Sci., Univ. of Iowa, Iowa City, IA, USA
IEEE Transactions on Medical Imaging (Impact Factor: 3.39). 07/2010; 29(6):1321 - 1330. DOI: 10.1109/TMI.2010.2047023
Source: IEEE Xplore


Optical coherence tomography (OCT) is becoming one of the most important modalities for the noninvasive assessment of retinal eye diseases. As the number of acquired OCT volumes increases, automating the OCT image analysis is becoming increasingly relevant. In this paper, a method for automated characterization of the normal macular appearance in spectral domain OCT (SD-OCT) volumes is reported together with a general approach for local retinal abnormality detection. Ten intraretinal layers are first automatically segmented and the 3-D image dataset flattened to remove motion-based artifacts. From the flattened OCT data, 23 features are extracted in each layer locally to characterize texture and thickness properties across the macula. The normal ranges of layer-specific feature variations have been derived from 13 SD-OCT volumes depicting normal retinas. Abnormalities are then detected by classifying the local differences between the normal appearance and the retinal measures in question. This approach was applied to determine footprints of fluid-filled regions-SEADs (Symptomatic Exudate-Associated Derangements)-in 78 SD-OCT volumes from 23 repeatedly imaged patients with choroidal neovascularization (CNV), intra-, and sub-retinal fluid and pigment epithelial detachment. The automated SEAD footprint detection method was validated against an independent standard obtained using an interactive 3-D SEAD segmentation approach. An area under the receiver-operating characteristic curve of 0.961 ? 0.012 was obtained for the classification of vertical, cross-layer, macular columns. A study performed on 12 pairs of OCT volumes obtained from the same eye on the same day shows that the repeatability of the automated method is comparable to that of the human experts. This work demonstrates that useful 3-D textural information can be extracted from SD-OCT scans and-together with an anatomical atlas of normal retinas-can be used for clinically important applications.

Full-text preview

Available from:
  • Source
    • "A significant improvement over autofluorescence patterns formulated for diagnosing fundus images can be achieved if continued cooperation among ophthalmic photographers, ophthalmologists and camera manufacturers to place this non-invasive, imaging modality at the forefront of ophthalmic imaging [5]. An automated method for 3-D analysis of retina texture, and quantification of fluid-filled regions (either intra-or subretinal fluid, or pigment epithelial detachment) associated with neovascular (or exudative) AMD termed as SEAD (Symptomatic Exudate-Associated Derangements) was proposed in [17] utilized for clinical settings in detecting retinal detachment. The translation of fundamental research findings in ophthalmology that will remain rapid in the future was projected in [1]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In today’s world, eye defects are more common among people of all ages. Most serious disorders of the eye include retinal detachment, macular degeneration. Technology is ever emergent and improving the way that assists in analyzing fundal images. Smart phones deployed with android applications leads to promising means for significant enrichment in eye care aids. In this paper an, ‘add-on’ for effective detection of such eye defects is presented that can be incorporated in smartphones. The developed add-on initially acquires L*a*b* triplets of the given fundus image. The resulting L*a*b* triplets is then contrast enhanced for further fundus examination as the image is captured under non-uniform illumination environment. Subsequent steps involve feature extraction and defect classification with Artificial neural network based Back Propagation Algorithm. Performance analysis of the proposed system is evaluated using fundus images attained from DRIVE, MESSIDOR and STARE database. The ROC analysis depicts a consistent performance and 90% classification accuracy for images of different database. Hence, this application improves efficacy of retinal diagnosis and aids in timely assessment of retinal disorders. This application is developed in android platform and is compatible with existing smartphones, augmenting its features. A step by step procedure for installation and operation of the add-on in smartphones is also presented in the paper.
    Full-text · Article · Nov 2014 · Biomedical sciences instrumentation
  • Source
    • "With bundle adjustment, the error becomes lower overall and is distributed more uniformly. 3) Multi-Surface Segmentation: For multi-surface segmentation , we build upon the method for segmenting 11 intra-retinal surfaces reported previously by our research group [31]. Due to large computational and memory requirements, the 11 intraretinal surfaces are not segmented simultaneously, but rather in a sequential and multiscale manner, in each step segmenting the surfaces lying in between the previously segmented ones [32]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: When segmenting intraretinal layers from multiple optical coherence tomography (OCT) images forming a mosaic or a set of repeated scans, it is attractive to exploit the additional information from the overlapping areas rather than discarding it as redundant, especially in low contrast and noisy images. However, it is currently not clear how to effectively combine the multiple information sources available in the areas of overlap. In this paper, we propose a novel graphtheoretic method for multi-surface multi-field co-segmentation of intraretinal layers, assuring consistent segmentation of the fields across the overlapped areas. After 2D en-face alignment, all the fields are segmented simultaneously, imposing a priori soft interfield-intrasurface constraints for each pair of overlapping fields. The constraints penalize deviations from the expected surface height differences, taken to be the depth-axis shifts that produce the maximum cross-correlation of pairwise-overlapped areas. The method's accuracy and reproducibility are evaluated qualitatively and quantitatively on 212 OCT images (20 9-field, 32 single-field acquisitions) from 26 patients with glaucoma. Qualitatively, the obtained thickness maps show no stitching artifacts, compared to pronounced stitches when the fields are segmented independently. Quantitatively, two ophthalmologists manually traced four intraretinal layers on ten patients, and the average error (4.581.46 m) was comparable to the average difference between the observers (5.861.72 m). Furthermore, we show the benefit of the proposed approach in co-segmenting longitudinal scans. As opposed to segmenting layers in each of the fields independently, the proposed co-segmentation method obtains consistent segmentations across the overlapped areas, producing accurate, reproducible, and artifact-free results.
    Full-text · Article · Jul 2014 · IEEE Transactions on Medical Imaging
  • Source
    • "Analysis of OCT images of the eye fundus is a technique known for many years which enables to acquire large amounts of information useful for medical diagnosis. Today, there are many known profiled algorithms for the analysis of both successive layers of the eye fundus [1-20] as well as the choroid layer [21]. Depending on the approach and practical implementation, these algorithms operate in the area of texture analysis [1,2], algorithms based on segmentation with Canny method [4], SVM [5], mathematical morphology [6], active contour [7,8], fuzzy algorithms [9], Markov model [10], robust segmentation [11], random contour analysis [12] and others [13-20]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In many practical aspects of ophthalmology, it is necessary to assess the severity degree of glaucoma in cases where, for various reasons, it is impossible to perform a visual field test - static perimetry. These are cases in which the visual field test result is not reliable, e.g. advanced AMD (Age-related Macular Degeneration). In these cases, there is a need to determine the severity of glaucoma, mainly on the basis of optic nerve head (ONH) and retinal nerve fibre layer (RNFL) structure. OCT is one of the diagnostic methods capable of analysing changes in both, ONH and RNFL in glaucoma.Material and method: OCT images of the eye fundus of 55 patients (110 eyes) were obtained from the SOCT Copernicus (Optopol Tech. SA, Zawiercie, Poland). The authors proposed a new method for automatic determination of the RNFL (retinal nerve fibre layer) and other parameters using: mathematical morphology and profiled segmentation based on morphometric information of the eye fundus. A quantitative ratio of the quality of the optic disk and RNFL - BGA (biomorphological glaucoma advancement) was also proposed. The obtained results were compared with the results obtained from a static perimeter. Correlations between the known parameters of the optic disk as well as those suggested by the authors and the results obtained from static perimetry were calculated. The result of correlation with the static perimetry was 0.78 for the existing methods of image analysis and 0.86 for the proposed method. Practical usefulness of the proposed ratio BGA and the impact of the three most important features on the result were assessed. The following results of correlation for the three proposed classes were obtained: cup/disk diameter 0.84, disk diameter 0.97 and the RNFL 1.0. Thus, analysis of the supposed visual field result in the case of glaucoma is possible based only on OCT images of the eye fundus. The calculations and analyses performed with the proposed algorithm and BGA ratio confirm that it is possible to calculate supposed mean defect (MD) of the visual field test based on OCT images of the eye fundus.
    Full-text · Article · Feb 2014 · BioMedical Engineering OnLine
Show more