Faculty Bibliography

BACKGROUND AND OBJECTIVE: Digital pathology has thus far focused on producing digital images of glass microscope slides. Spectral domain optical coherence tomography (SD-OCT) can be used to directly scan tissue blocks to produce three-dimensional histology images, potentially bypassing glass slide workflow. MATERIALS AND METHODS: Formalin-fixed paraffin-embedded tissue blocks were scanned using SD-OCT and resulting images were compared with corresponding areas on microscope slides. RESULTS: Low-magnification features were recognizable, including tissue outlines, fat, vessels, and outlines of colonic mucosal crypts. Subtle textures that were suggestive of benign breast lobules and ovarian tumor features were also visible. Initial SD-OCT images lacked resolution and contrast relative to traditional microscopy, but the image content suggests that additional features of interest are present and may be revealed with improved SD-OCT resolution and more post-processing experience. Elucidation of three-dimensional histology and pathology are also future tasks. CONCLUSION: Eventual availability of diagnostic-quality three-dimensional histology would have a profound impact on anatomic pathology.

Three dimensional (3D) ophthalmic imaging using optical coherence tomography (OCT) has revolutionized assessment of the eye, the retina in particular. Recent technological improvements have made the acquisition of 3D-OCT datasets feasible. However, while volumetric data can improve disease diagnosis and follow-up, novel image analysis techniques are now necessary in order to process the dense 3D-OCT dataset. Fundamental software improvements include methods for correcting subject eye motion, segmenting structures or volumes of interest, extracting relevant data post hoc and signal averaging to improve delineation of retinal layers. In addition, innovative methods for image display, such as C-mode sectioning, provide a unique viewing perspective and may improve interpretation of OCT images of pathologic structures. While all of these methods are being developed, most remain in an immature state. This review describes the current status of 3D-OCT scanning and interpretation, and discusses the need for standardization of clinical protocols as well as the potential benefits of 3D-OCT scanning that could come when software methods for fully exploiting these rich datasets are available clinically. The implications of new image analysis approaches include improved reproducibility of measurements garnered from 3D-OCT, which may then help improve disease discrimination and progression detection. In addition, 3D-OCT offers the potential for preoperative surgical planning and intraoperative surgical guidance.

PURPOSE: Measurements of human Schlemm's canal (SC) have been limited to histologic sections. The purpose of this study was to demonstrate noninvasive measurements of aqueous outflow (AO) structures in the human eye, examining regional variation in cross-sectional SC areas (on/off collector channel [CC] ostia [SC/CC] and nasal/temporal) in the eyes of living humans. METHODS: SC was imaged by spectral-domain optical coherence tomography with a 200-nm bandwidth light source. Both eyes of 21 healthy subjects and one glaucomatous eye of three subjects were imaged nasally and temporally. Contrast and magnification were adjusted to maximize visualization. Cross-sectional SC on and off SC/CC was traced three times by two independent masked observers using ImageJ (ImageJ 1.40g, http://rsb.info.nih.gov/ij/ Wayne Rasband, developer, National Institutes of Health, Bethesda, MD). The mean SC area was recorded. A linear mixed-effects model was used to analyze eye, nasal/temporal laterality, and SC area on or off SC/CC. RESULTS: SC area was significantly larger on SC/CCs than off (12,890 vs. 7,391 micorm(2), P < 0.0001) and was significantly larger on the nasal side than on the temporal (10,983 vs. 8,308 micorm(2), P = 0.009). SC areas were significantly smaller in glaucoma patients than in normal subjects, whether pooled (P = 0.0073) or grouped by on (P = 0.0215) or off (P = 0.0114) SC/CC. CONCLUSIONS: Aqueous outflow structures, including SC and CCs, can be noninvasively assessed in the human eye. These measurements will be useful in physiological studies of AO and will be clinically useful in the determination of the impact of glaucoma therapies on IOP as well as presurgical planning.

PURPOSE: To develop a fully automated algorithm (AP) to perform a volumetric measure of the optic disc using conventional stereoscopic optic nerve head (ONH) photographs, and to compare algorithm-produced parameters with manual photogrammetry (MP), scanning laser ophthalmoscope (SLO) and optical coherence tomography (OCT) measurements. METHODS: One hundred twenty-two stereoscopic optic disc photographs (61 subjects) were analyzed. Disc area, rim area, cup area, cup/disc area ratio, vertical cup/disc ratio, rim volume and cup volume were automatically computed by the algorithm. Latent variable measurement error models were used to assess measurement reproducibility for the four techniques. RESULTS: AP had better reproducibility for disc area and cup volume and worse reproducibility for cup/disc area ratio and vertical cup/disc ratio, when the measurements were compared to the MP, SLO and OCT methods. CONCLUSION: AP provides a useful technique for an objective quantitative assessment of 3D ONH structures.

PURPOSE: Time domain optical coherence tomography (TD-OCT) has been used commonly in clinical practice, producing a large inventory of circular scan data for retinal nerve fiber layer (RNFL) assessment. Spectral domain (SD)-OCT produces three-dimensional (3-D) data volumes. The purpose of this study was to create a robust technique that makes TD-OCT circular scan RNFL thickness measurements comparable with those from 3-D SD-OCT volumes. METHODS: Eleven eyes of 11 healthy subjects and 7 eyes of 7 subjects with glaucoma were enrolled. Each eye was scanned with one centered and eight displaced TD-OCT scanning circles. One 3-D SD-OCT cube scan was obtained at the same visit. The matching location of the TD-OCT scanning circle was automatically detected within the corresponding 3-D SD-OCT scan. Algorithm performance was assessed by estimating the difference between the detected scanning circle location on 3-D SD-OCT volume and the TD-OCT circle location. Global and sectoral RNFL thickness measurement errors between the two devices were also compared. RESULTS: The difference (95% confidence interval) in scanning circle center locations between TD- and SD-OCT was 2.3 (1.5-3.2) pixels (69.0 [45.0-96.0] microm on the retina) for healthy eyes and 3.1 (2.0-4.1) pixels (93.0 [60.0-123.0] microm on the retina) for glaucomatous eyes. The absolute RNFL thickness measurement difference was significantly smaller with the matched scanning circle. CONCLUSIONS: Scan location matching may bridge the gap in RNFL thickness measurements between TD-OCT circular scan data and 3-D SD-OCT scan data, providing follow-up comparability across the two generations of OCTs.

Three-dimensional optical coherence tomography (OCT) is a new ophthalmic imaging technique offering more detailed quantitative analysis of the retinal structure. Eye movement during 3D OCT scanning, however, creates significant spatial distortions that may adversely affect image interpretation and analysis. Current software solutions must use additional reference images or B-scans to correct eye movement in a certain direction. The proposed particle filtering algorithm is an independent 3D alignment approach, which does not rely on any reference image. 3D OCT data is considered as a dynamic system, while location of A-scan is represented by the state space. A particle set is generated to approximate the probability density of the state. The state of the system is updated frame by frame to detect A-scan movement. Seventy-four 3D OCT images with eye movement were tested and subjectively evaluated by comparing them with the original images. All the images were improved after z-alignment, while 81.1% images were improved after x-alignment. The proposed algorithm is an efficient way to align 3D OCT volume data and correct the eye movement without using references.

We address a novel problem domain in the analysis of optical coherence tomography (OCT) images: the diagnosis of multiple macular pathologies in retinal OCT images. The goal is to identify the presence of normal macula and each of three types of macular pathologies, namely, macular hole, macular edema, and age-related macular degeneration, in the OCT slice centered at the fovea. We use a machine learning approach based on global image descriptors formed from a multi-scale spatial pyramid. Our local descriptors are dimension-reduced Local Binary Pattern histograms, which are capable of encoding texture information from OCT images of the retina. Our representation operates at multiple spatial scales and granularities, leading to robust performance. We use 2-class Support Vector Machine classifiers to identify the presence of normal macula and each of the three pathologies. We conducted extensive experiments on a large dataset consisting of 326 OCT scans from 136 patients. The results show that the proposed method is very effective.