Retina Today is pleased to introduce the third installment of a new column, Global Perspectives. With the international reach of the publication, the editors seek to provide insight into techniques used by retina surgeons around the world, and to offer a global perspective on patient-management issues and treatment strategies. We look forward to hearing your ideas for future columns. -Albert J. Augustin, MD; Stanislao Rizzo, MD; and J. Fernando Arevalo, FACS

A new generation of optical coherence tomography (OCT) machines, termed spectral-domain (SD) or, less frequently, Fourier-domain (FD), arrived on the market in 2007 and are now proposed by at least six companies in Europe. SD OCT and time-domain (TD) OCT (Stratus; Carl Zeiss Meditec, Dublin CA) are based on the same principle of interferometry, which allows retinal scan profiles to be acquired and the main intraretinal layers to be distinguished. SD OCT sets itself apart from the TD generations via its outstanding sensitivity, allowing high-speed 2-D and 3-D imaging of weak backscattering retinal tissue with high axial resolution. The signals in SD OCT are presented in frequency space, and the method of obtaining and analyzing the light signal from the sample differs from TD OCT. For example, the Stratus OCT performs two scans over the sample: an axial-depth scan (A-scan) by moving the reference mirror and a lateral scan (B-scan) by moving the beam on the sample. In SD OCT the A-scan is replaced by a spectroscopic signal analysis that yields a whole A-scan trace without scanning. Information from the depth scan is provided by an inverse Fourier transform of the spectrum of the back-scattered light. Only the lateral scan has to be performed in an SD OCT, which considerably reduces the acquisition time.

FASTER SPEED, HIGHER RESOLUTION
The main advantage to SD OCT technology is the speed of data acquisition, which facilitates an increased number of lines that scan the retina. Instead of the six radial lines of 512 A-scans, SD OCT is able to record at least 128 lines of 512 A-scans covering a surface of 6x6 mm in about 2 sec (ie, approximately the same time as the Stratus). Consequently, it is now possible to have a 3-D detailed map of the posterior pole. It also may record fewer lines but with a high longitudinal resolution of 1024 to 4096 A-scans by 6 mm.

The axial resolution has also been improved and is now 5 to 7 µm, compared with 10 µm with the Stratus, resulting in better visualization of the different intraretinal layers and offering the possibility of a delineation (segmentation) of the retinal pigment epithelium (RPE) and the internal limiting membrane (ILM), and consequently the volume comprised between these two borders. Theoretically, new algorithms should be able to delineate more than two retinal layers. In pathologic cases, however, the presence of artifacts often impairs the segmentation, and only the planes of RPE and ILM are robust enough to be detected and mapped.

RELIABILITY IMPROVED
Another advantage of SD OCT is its acquisition mode. Because the SD OCT automatically scans lines within a 6x6 mm box, most of the lesions that are be present in the posterior pole are detected, and the acquisition of the SD OCT map is less dependent on the operator.

The reliability of macular thickness calculation of various models of SD OCT has been tested comparatively with the Stratus OCT and appears to be good for all models. However, the normal value of retinal thickness is greater than with the Stratus. This is due to the fact that the Stratus OCT uses the inner segment/outer segment photoreceptor (IS/OS PR) line as the reference for the outer retina boundary, which is incorrect. SD OCT technology has corrected this error by placing the outer boundary of the retina at the level of the RPE. The retinal thickness measurements with SD OCT are now 40 to 60 µm thicker than with the Stratus. The thickness varies between the center and the periphery of the macula and also according to the machine being used.

ALL IN THE DETAILS
The most important issue when evaluating SD OCT technology is this: When compared with the Stratus, do the images obtained with SD OCT show more relevant details, help obtain a more accurate diagnosis, or guide treatment more effectively? To answer this question, one should utilize comparable Stratus and SD OCT modes. For instance, Stratus fast map at 128 A-scans per line should not be compared with a high-definition scan with SD OCT comprising 4096 A-scans per line. A good comparison, rather, would be between the high-definition map of the Stratus (six lines of 512 A-scans) and an SD OCT map made of 128 lines of 512 A-scans. In this case the longitudinal resolution of the scans is identical; however, due to better axial resolution each scan is slightly more precise than with the Stratus. The map is also more accurate because the calculation of retinal thickness at each point requires much less interpolation.

The possibility offered by the different SD OCT devices to acquire high-definition scans with longitudinal resolution of 1024 to 4096 A-scans also provides images much more precise than those of the Stratus. Some devices offer the additional possibility to average several individual scan lines, either in real time or after acquisition; this mode decreases the background speckle and enhances the transition between intraretinal layers. It may, however, also be the cause of new artifacts.

In practice, SD OCT improves the visibility of outer retinal structures, including four outer retinal bands: (1) the ILM, which is the junction between the apex of MŸller cells and photoreceptor inner segments; (2) the junction between the inner and outer segments of the photoreceptors, which contains a refractile body; (3) a line that is supposed to be the tip of the outer segments of the cones, which are typically shorter than the rods; and (4) the outermost, which is the RPE band.

Spectral-domain OCT also improves the visibility of the vitreous cortex, which appears more reflective and thicker than with the Stratus.

SD OCT FOR EVERYDAY PRACTICE?
In pathologic cases, SD OCT has several advantages over TD OCT. SD OCT presents less risk in forgetting a small lesion than with the six radial scans of the Stratus. Additionally, the details of intraretinal anomalies are more accurate than with the Stratus. In age-related macular degeneration, for instance, it is easier to document a retinal angiomatous proliferation that could have been missed between two radial lines of the Stratus. In cystoid macular edema, the intraretinal cysts appear to be more numerous, more tiny subfoveal serous retinal detachments are be disclosed, and the IS/OS PR line may appear continuous with SD OCT, where it may seem more broken with the Stratus. In diabetic macular edema (DME), the SD OCT mapping is also more informative than with the Stratus and is better at delineating the areas of retinal thickening and their location regarding the foveal center. Moreover, further improvements in the software of the SD OCT devices will probably improve their calculation capabilities and reduce the artifacts that continue to exist, particularly in the determination of the outer borders of the retina.

For everyday clinical use, however, it is not certain that the SD OCT will take a qualitative leap forward. We view SD OCT, rather, as progress that makes it easier to record reproducible maps and detect intraretinal changes. It is important to remember that the only anatomic parameters we look at to treat or retreat choroidal neovascularization are the presence of an RPE detachment, subretinal fluid or fibrosis, and retinal thickening. Although these data are well shown with SD OCT, they were rarely missed by the Stratus. For DME, the parameters are even simpler: central macular thickness, and possibly the presence of subfoveal fluid—easily distinguishable on the Stratus.

The advent of SD OCT shows that the OCT technology is quickly evolving. Improvements will continue to debut. OCT resolution will be improved at a cellular level by coupling with adaptive optics and will also be able to measure the retinal blood flow in small vessels by Doppler effect or to record the focal electric activity of small retinal areas.

Alain Gaudric, MD, is Professor and Head of the Department of Ophthalmology, Hopital Lariboisière, Université Paris. He states that he has no financial interests to disclose. Dr. Gaudric can be reached at: alain.gaudric@lrb.aphp.fr.