Summary: Dr. Cabrera Fernández's research focuses on physical and mathematical modeling of retinal morphology as visualized by Optical Coherence Tomography in order to quantify treatment-induced changes in patients with ocular diseases.

Delia Cabrera Fernandez, Ph.D.
Research Assistant Professor

View published research articles by this doctor in the National Library of Medicine and in Conference Proceedings.

Current Research Summary: Optical Coherence Tomography (OCT) has developed rapidly since its potential for application in clinical medicine was first demonstrated in 1991. An OCT image represents a cross-sectional, micron scale picture of the optical reflectance properties of the tissue. This image can either be used to qualitatively assess retinal features and pathologies or to objectively make quantitative measurements. In order to improve the quantitative analysis of the commercial STRATUSOCT system, Dr. Cabrera Fernández has been working in tthe implementation of new methodologies to extract more information from the retinal images. Two particular approaches are currently under development.

I. Segmentation of retinal layers

Segmentation of retinal layers from OCT images is fundamental to diagnose the progress of retinal diseases. We have recently shown that the retinal layers can be automatically and/or interactively located with good accuracy with the aid of local coherence information of the retinal structure. OCT images are processed using the ideas of texture analysis by means of the structure tensor combined with complex diffusion filtering. Experimental results indicate that our proposed novel approach has good performance in speckle noise removal, enhancement and segmentation of the various cellular layers of the retina using the commercial STRATUSOCT unit. The following image shows the segmentation results obtained for a normal subject.

Once the various cellular layers of the retina are segmented the thickness mapping for specific layer structures could be obtained:

The methodology under development could have an important role in the future development of computer-assisted OCT quantification techniques. For example, the thickness and volume of the retinal layers, and the reflectance profile varying with depth could be obtained to provide a more complete OCT biometry for diagnosis.

II. Delineating fluid-filled region boundaries in retinal OCT images

Calculation of retinal thickness and volume by the current OCT system includes fluid-filled regions along with actual retinal tissue.  In order to quantify these areas independently from the retinal tissue, they must be outlined.

A typical example of OCT images of the human macular retina for normal and pathologic eyes is shown in Figure 1A and B, respectively. 

The OCT image shown in Figure 1B is from a subject with neovascular age-related macular degeneration and cystoid macular edema. This image demonstrates thickening of the macula with several large hyporeflective cystoids spaces in the fovea. In comparing the image of the pathologic subject with the one obtained in the normal subject, the importance of quantifying the structural changes of retinal features and pathologies is more than obvious. However, the commercially available STRATUSOCT that is currently used in the standard clinical practice is only able to provide a qualitative, visual assessment of retinal features. For example, the STRATUSOCT boundary-detection system outlines the inner and outer boundaries of the retina to provide a final measure of retinal thickness, but the STRATUSOCT software analysis does not provide a quantitative measure of the pathologic features (e.g., cystoid volume) which may aid in assessing the progression of disease and the success of therapy in a more complete and quantitative approach.

Using an edge-based segmentation technique, Dr. Cabrera Fernández has evaluated the ability of an active contour model to yield accurate shape descriptions of fluid-filled regions associated with neovascular age-related macular degeneration (ARAMD). An active contour model is a computer generated curve that move within images to find object boundaries. Figure 2 shows the performance of the active contour model on a set of six radial OCT scans obtained from an AMD patient having multiple lesions in the central part of the retina. The fluid-filled region boundaries are outlined in red. The outer and inner layers of the retina have been outlined in yellow.

Since the number and size of fluid-filled regions indicate the severity of the disease in the patient, fluid-filled region detection may significantly aid in analysis of treatments and diagnosis. With such quantification, the accuracy and precision of OCT will allow improved monitoring of patients, earlier detection of pathology, and more precise treatment protocols. The detection method tested is demonstrably effective in capturing the complexity of fluid-filled regions in OCT images, allowing us to quantify these fluid-filled regions and subretinal fluid areas independently from the retinal tissue.

Dr. Cabrera Fernández’s long-term goal is to improve the ability of OCT to correlate changes in retinal morphology with changes in visual acuity after therapy.