Chiu SJ, Li XT, Nicholas P, et al.. Automatic segmentation of seven retinal layers in SDOCT images congruent with expert manual segmentation

Department of Biomedical Engineering, Duke University, Durham, NC 27708, USA.
Optics Express (Impact Factor: 3.49). 08/2010; 18(18):19413-28. DOI: 10.1364/OE.18.019413
Source: PubMed

ABSTRACT Segmentation of anatomical and pathological structures in ophthalmic images is crucial for the diagnosis and study of ocular diseases. However, manual segmentation is often a time-consuming and subjective process. This paper presents an automatic approach for segmenting retinal layers in Spectral Domain Optical Coherence Tomography images using graph theory and dynamic programming. Results show that this method accurately segments eight retinal layer boundaries in normal adult eyes more closely to an expert grader as compared to a second expert grader.

136 Reads
  • Source
    • "Within a given eye of an individual animal, thirty B-scans were averaged to reduce the speckle noise in the image, corresponding to lateral averaging over 10 lm. A semi-automated segmentation software developed at Duke University (Chiu et al., 2010; Lee et al., 2013) was used to calculate the thickness and average intensity of each layer. These measurements were made sequentially at the same eccentricity in the same animal over time, resulting in longitudinal data. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Microglia dynamically prune synaptic contacts during development, and digest waste that accumulates in degeneration and aging. In many neurodegenerative diseases, microglial activation and phagocytosis gradually increase over months or years, with poorly defined initial triggering events. Here, we describe rapid retinal microglial activation in response to physiological light levels in a mouse model of photoreceptor degeneration that arises from defective rhodopsin deactivation and prolonged signaling. Activation, migration and proliferation of microglia proceeded along a well-defined time course apparent within 12h of light onset. Retinal imaging in vivo with optical coherence tomography (OCT) revealed dramatic increases in light-scattering from photoreceptors prior to the outer nuclear layer thinning classically used as a measure of retinal neurodegeneration. This model is valuable for mechanistic studies of microglial activation in a well-defined and optically accessible neural circuit, and for the development of novel methods for detecting early signs of pending neurodegeneration in vivo.
    Vision Research 09/2014; 102. DOI:10.1016/j.visres.2014.07.011 · 1.82 Impact Factor
  • Source
    • "We have previously demonstrated novel algorithms for pathology segmentation that may be applied to identify the rapid alterations that occur during surgical intervention [4], [39]. Automated segmentation of retinal layers has been previously described [40]. The application of similar algorithms could potentially be utilized for real-time surgical feedback on instrument-tissue proximity. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Purpose To demonstrate key integrative advances in microscope-integrated intraoperative optical coherence tomography (iOCT) technology that will facilitate adoption and utilization during ophthalmic surgery. Methods We developed a second-generation prototype microscope-integrated iOCT system that interfaces directly with a standard ophthalmic surgical microscope. Novel features for improved design and functionality included improved profile and ergonomics, as well as a tunable lens system for optimized image quality and heads-up display (HUD) system for surgeon feedback. Novel material testing was performed for potential suitability for OCT-compatible instrumentation based on light scattering and transmission characteristics. Prototype surgical instruments were developed based on material testing and tested using the microscope-integrated iOCT system. Several surgical maneuvers were performed and imaged, and surgical motion visualization was evaluated with a unique scanning and image processing protocol. Results High-resolution images were successfully obtained with the microscope-integrated iOCT system with HUD feedback. Six semi-transparent materials were characterized to determine their attenuation coefficients and scatter density with an 830 nm OCT light source. Based on these optical properties, polycarbonate was selected as a material substrate for prototype instrument construction. A surgical pick, retinal forceps, and corneal needle were constructed with semi-transparent materials. Excellent visualization of both the underlying tissues and surgical instrument were achieved on OCT cross-section. Using model eyes, various surgical maneuvers were visualized, including membrane peeling, vessel manipulation, cannulation of the subretinal space, subretinal intraocular foreign body removal, and corneal penetration. Conclusions Significant iterative improvements in integrative technology related to iOCT and ophthalmic surgery are demonstrated.
    PLoS ONE 08/2014; 9(8):e105224. DOI:10.1371/journal.pone.0105224 · 3.23 Impact Factor
  • Source
    • "Based on these considerations, layer segmentation was performed using an algorithm based on Chiu et al. [19] which segmented the positions of inner limiting membrane (ILM) inner segments (IS) and retinal pigment epithelium (RPE) and based on these positions, retinal thickness (defined to be RPE – ILM). In order to use additional knowledge and improve robustness, Chiu et al.’s algorithm was extended so that multiple layers could be found in a single Dijkstra shortest path search. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Variability in illumination, signal quality, tilt and the amount of motion pose challenges for post-processing based 3D-OCT motion correction algorithms. We present an advanced 3D-OCT motion correction algorithm using image registration and orthogonal raster scan patterns aimed at addressing these challenges. An intensity similarity measure using the pseudo Huber norm and a regularization scheme based on a pseudo L0.5 norm are introduced. A two-stage registration approach was developed. In the first stage, only axial motion and axial tilt are coarsely corrected. This result is then used as the starting point for a second stage full optimization. In preprocessing, a bias field estimation based approach to correct illumination differences in the input volumes is employed. Quantitative evaluation was performed using a large set of data acquired from 73 healthy and glaucomatous eyes using SD-OCT systems. OCT volumes of both the optic nerve head and the macula region acquired with three independent orthogonal volume pairs for each location were used to assess reproducibility. The advanced motion correction algorithm using the techniques presented in this paper was compared to a basic algorithm corresponding to an earlier version and to performing no motion correction. Errors in segmentation-based measures such as layer positions, retinal and nerve fiber thickness, as well as the blood vessel pattern were evaluated. The quantitative results consistently show that reproducibility is improved considerably by using the advanced algorithm, which also significantly outperforms the basic algorithm. The mean of the mean absolute retinal thickness difference over all data was 9.9 um without motion correction, 7.1 um using the basic algorithm and 5.0 um using the advanced algorithm. Similarly, the blood vessel likelihood map error is reduced to 69% of the uncorrected error for the basic and to 47% of the uncorrected error for the advanced algorithm. These results demonstrate that our advanced motion correction algorithm has the potential to improve the reliability of quantitative measurements derived from 3D-OCT data substantially.
    Biomedical Optics Express 08/2014; 5(8). DOI:10.1364/BOE.5.002591 · 3.65 Impact Factor
Show more