OCT of the Macula

An expert provides a primer on useful scans, identifying artifacts and time domain vs. spectral domain technology.

By Glenn J. Jaffe, MD

While OCT continues to expand our knowledge of retinal pathology and becomes an increasingly important tool in clinical trials, in the OR and in anterior segment evaluation, it is primarily used in vitreoretinal practice to image the macula. By allowing us to obtain morphologic, i.e., anatomic information as well as quantitative thickness measurements, the technology provides valuable information to identify various disease states, determine the cause of decreased vision and monitor treatment.

With the cross-sectional images captured by OCT we can visualize morphologic features such as epiretinal membranes, intraretinal, subretinal and sub-retinal pigment epithelium (RPE) fluid, and vitreoretinal interface changes, including vitreomacular adhesions (VMAs) and vitreomacular traction (VMT). We can also obtain information about abnormal tissue beneath the retina, especially in the choroid.

The ability to view the retinal anatomy in this manner helps us to not only identify pathology, but also to determine the anatomic factors that contribute to vision loss. As such, the cross-sectional OCT images improve our ability to choose the appropriate treatment and monitor its effectiveness in the clinic. For example, OCT can tell us whether the cause of decreased visual acuity in a patient with age-related macular degeneration (AMD) is progression from the dry to the wet form of the disease or the development of vitreomacular traction. The treatment for each, of course, would be entirely different. If wet AMD is the diagnosis, anti-VEGF injections would be the likely therapy. For vitreomacular traction, surgery may be necessary, or, in the near future, we might inject a vitreous pharmacolysis agent. In addition, when we are treating macular edema associated with various diseases, we can see how well we are doing based on the qualitative changes in fluid observed in follow-up images.

Just as useful as the morphologic information OCT provides is its quantitative measurement of retinal thickness, which is presented in numeric thickness maps. This quantitative data is

now frequently used in clinical trials as criteria for eligibility and retreatment. Many landmark trials have based their endpoints on macular, i.e., central subfield, thickness as measured by OCT. In clinical practice as well, the quantitative values are especially useful for monitoring change over time.

OCT-RELATED CONSIDERATIONS FOR TODAY’S PRACTICE

As OCT technology evolves, questions arise as to the best way to use these improvements most effectively and efficiently in our practices. Two issues we must consider are which of the many available scanning patterns we should rely on routinely to obtain the most useful information, and whether it is necessary to switch from time-domain OCT (TDOCT) to newer generation spectral-domain OCT (SDOCT). In addition, image artifacts may be less frequent with SDOCT than with TDOCT due to spectral domain’s faster acquisition speed, improved signal to noise ratio, image averaging, and improved algorithms, but they do occur with both types of devices.¹ Knowing how to recognize them is key to assure that the quantitative thickness measurements are accurate.

PREFERRED MACULAR SCANS WITH TDOCT AND SDOCT

TDOCT remains relevant in clinical practice. Practices that have invested in an SDOCT device often continue to use their TDOCT unit(s). We can be confident that the qualitative information and the quantitative measurements of thickness changes over time obtained from TDOCT allow us to provide quality care for the majority of conditions we regularly see. TDOCT was used very effectively in the recent clinical trial of Ocriplasmin (ThromboGenics) for symptomatic VMA, which was recommended for approval by the FDA’s Dermatologic and Ophthalmic Drugs Advisory Committee in July. Also, based on the TDOCT-SDOCT comparison that was performed in CATT (Comparison of Age-Related Macular Degeneration Treatments Trials), study results would not have differed significantly had the investigators used SDOCT rather than TDOCT to guide their treatment decisions.²

In my clinic, I prefer to obtain a high-resolution scan as well as a multiple-line scan that results in a map of a larger area, which I call “volume scanning,” for each patient. Using both provides the qualitative/morphologic and the quantitative information I need.

Therefore, if I am in a clinic where only TDOCT is available, I obtain the Macular Thickness Map and the Fast Macular Thickness Map protocols. Both scan patterns provide quantitative data and both are composed of six radial scans oriented 30° apart that cover a 6-mm area. The Fast Macular Thickness Map is lower resolution than the Macular Thickness Map, but it is helpful because thickness measurements are quicker, easier to obtain and less dependent on patient fixation. Often, with TDOCT, I also order a 7-mm scan, which starts at the midpoint of the optic nerve and is offset approximately 5° from horizontal. This results in a high-resolution scan close to the foveal center that is also less dependent on good patient fixation.

If I am using SDOCT, I order a high-resolution raster scan. With the Cirrus HD-OCT (Carl Zeiss Meditec), this is typically a five-line raster scan. For the Spectralis (Heidelberg Engineering), it is a 7-line scan. These are particularly helpful for visualizing the fine details of macular pathology. With SD-OCT, I also order a volume cube. Depending on the brand of machine, the number of scans that make up the cube differs, but all of them are an excellent way to track macular thickness changes. They also show macular morphology, typically over a 6-mm by 6-mm area.

With either TDOCT or SDOCT, if I suspect a particular pathology, and the area is not adequately covered by the standard scanning protocols, I request that the technician move the scanner to specifically image the area of interest. This can be helpful, for example, for differentiating a retinal detachment from a retinal schisis.

RECOGNIZING THE PRESENCE OF IMAGE ARTIFACTS

As previously mentioned, image artifacts can occur with either TDOCT or SDOCT and can lead to a misrepresentation of macular thickness. Of the types of artifacts that can occur, those caused by the segmentation algorithm’s misidentification of the retinal boundaries, i.e., the internal limiting membrane (ILM) and the RPE, decentration and image degradation are most significantly associated with thickness calculation errors. Therefore, it is crucial to be alert for these artifacts.

Several abnormalities, including epiretinal membranes and vitreous detachment, can lead to incorrect identification of the ILM. Subretinal neovascular lesions are a common cause of an improperly identified RPE. A good indication of a boundary error artifact is unusually shaped thick or thin areas with sharp margins on the pseudocolor surface map. When this is seen, the B-scan image should be checked to see if the inner and outer retinal boundaries are misplaced. If they are, both the Stratus TDOCT (Carl Zeiss Meditec) and the SDOCT instruments enable manual correction.

Decentration artifacts are often operator-dependent, but they can also be caused by morphology that prevents identification of the fovea, or poor patient fixation. If the fovea as represented by the pseudocolor surface map does not correspond to the central subfield on the map’s grid, a centration error should be suspected. If the macula is significantly thickened, for example by cystoid macular edema (CME), a foveal depression may not be visible. In a case such as this, because we would expect the CME to be centered on the fovea, the CME should be centered on the pseudocolor map. If not, a decentration artifact is likely present. Fortunately, with our SDOCT devices, we can adjust the grid after the fact to center the image. However, when that is done, data at the edges of the grid are lost, which affects the accuracy of the total macular volume calculation. If a foveal depression is not visible, we can also use some anatomical clues on the cross-sectional images that would indicate a decentered image. The retinal nerve fiber layer becomes very thin in the center, and the outer nuclear layer, also called outer nuclear complex, is thicker in the center.

Image degradation, which can be caused by media opacity (e.g., corneal opacity, cataract or vitreous opacity) or dry eye, is another source of artifacts. Having patients with dry eye blink before the scan, or instilling artificial tears, can help to minimize this problem, as the computer will be better able to identify the retinal boundaries. In some cases of media opacity, there may be little we can do to capture quality images. However, in many cases, useful information, especially from the B-scan images, can be obtained despite a cataract or other media opacity. Often we can still see the inner and outer retina, which means we can adjust the boundaries to obtain thickness information if doing so is crucial.

THE PROMISE OF SDOCT AND BEYOND

While TDOCT and SDOCT continue to coexist, within the past 2 years, we have seen a significant shift to the newer generation of the technology. Through research and experience, we are becoming more familiar with how its advantages — including faster image acquisition, improved resolution and the ability to register images from one patient visit to another — may benefit physcians and our patients. We are finding that SDOCT can provide additional information that is difficult or impossible to obtain with TDOCT. For example, in general, SDOCT gives a better view of epiretinal membranes. It also provides better visualization of outer retinal structures, particularly the external limiting membrane and the inner segment ellipsoid line (inner segment/outer segment junction). Emerging evidence suggests that the integrity of those structures may correlate more closely than macular thickness with visual acuity. Knowing their status, therefore, could be helpful in predicting how much vision a patient could recover after treatment. SDOCT also provides much more detailed information about the choroidal blood vessels.

Further advances that will enhance our analysis of the macula are on the way. Portable, even handheld, devices are already in use in operating rooms and newborn nurseries. Intraoperative OCT has great potential for helping us improve surgical outcomes. The technology will likely be miniaturized even further, perhaps making it possible for patients to image themselves at home or, for example, at kiosks in malls.

Machines capable of even higher resolution than we use in practice today have already been developed. Also in development are approaches to OCT based on wavelength swept lasers (swept-source OCT), which achieve faster scanning and higher resolution that may better allow us to image the retina and choroid.

Another approach, based on Doppler technology, may allow us to measure blood flow, which could significantly improve our evaluation and management of retinal vascular diseases. Also on the horizon is the use of adaptive optics as a way to compensate for optical artifacts and potentially provide high-resolution images of rods and cones.

Dr. Jaffe is a vitreoretinal surgeon, professor of Ophthalmology and chief of the Vitreoretinal Diseases and Surgery Service at Duke University Eye Center in Durham, N.C. He is also the founder and director of the Duke Reading Center.

References

1. Han IC, Jaffe GJ. Ophthalmology. Evaluation of artifacts associated with macular spectral-domain optical coherence tomography. Ophthalmology 2010;117(6):1177-1189.
2. Comparison of Age-Related Macular Degeneration Treatments Trials (CATT) Research Group. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: two-year results. Ophthalmology 2012;119(7):1388-1398.