SYMPOSIUM RECAP

Exploring New Dimensions in OCT Imaging

Spectral domain OCT offers clinicians new perspectives in visualizing retinal disease.

By Peter K. Kaiser, M.D.

When optical coherence tomography (OCT) was introduced in the early 1990s, many physicians questioned the clinical utility of the technology. Some thought it would be useful for research, but nothing more. Since then, OCT imaging has become an integral part of treating retinal and glaucomatous diseases.

Time domain OCT revolutionized ophthalmology practices, but top researchers and physicians continued to investigate ways to improve the technology. The result of this continued exploration is spectral domain OCT, also known as Fourier domain OCT. Here, I'll discuss how spectral domain OCT quickly generates precise images that help us better diagnose, monitor and treat disease.

Increased Speed

The key difference between time domain and spectral domain OCT is A-scan acquisition speed. In both cases, the light reflected from the patient is interfered with light reflected within a reference arm. To see an interference pattern that can be reconstructed into an A-scan profile of the patient's tissue, the distance traveled by light in the reference arm must be the same as the round-trip distance to the tissue of interest.

In time domain OCT, an A-scan is accomplished by changing the length of the reference arm. This allows detection of the reflection profile corresponding to a range of depths. As the depth range is scanned, the interference signal is digitized and processed to generate the OCT images with which we've become familiar. A typical time domain B-scan involves 512 A-scans captured over approximately 1 second.

With spectral domain technology, the interference from a range of depths is captured almost instantaneously by taking a snapshot of the light's spectral composition. High-speed digital signal processing converts the interference spectrum into an A-scan. Spectral domain works well at up to 30,000 A-scans per second with signal-to-noise ratios comparable to the time domain method. Thus, many B-scans can be acquired on a 1- to 2-second time scale. This minimizes motion artifacts and allows entire 3-D volumes to be scanned in the time previously required for a single B-scan.

Furthermore, because spectral domain OCT is so rapid, we can image the macula with a 6 mm × 6 mm box and every infinitesimal point within that box can be analyzed. This isn't possible with time domain OCT, since the amount of patient movement during the time it takes to image the same area makes the resulting information useless. Spectral domain OCT allows us to measure 100% of the retinal thickness points within this box vs. about 5% of the points with time domain OCT (using a radial line scan). Thus, with spectral domain OCT, we can detect very small lesions that might otherwise be missed.

Superior Resolution

There are two independent directions for resolution in OCT. One is lateral (or transverse) resolution, which depends on the optics of the eye and thus is difficult to upgrade. Axial (or depth) resolution, however, is based on the wavelength of the light source used in the machine, so we can improve the axial resolution by changing our light source.

Time domain OCT uses a superluminescent diode, broadband light source and offers 8- to 10-micron resolution. Spectral domain OCT uses an improved superluminescent diode, so we can achieve about 5- to 7-micron resolution. If you want even higher axial resolution, you could purchase a special laser light source, which provides a very broad bandwidth and approximately 2-micron resolution — the theoretical limit of what can be achieved with OCT technology. However, these lasers are very expensive and currently are used only in research labs.

With improved axial resolution and denser scanning, we can see considerably more detail with spectral domain OCT than time domain systems. For instance, it's sometimes difficult to distinguish the external limiting membrane on time domain OCT, but it's very easy to discern on spectral domain systems, such as the Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, Calif.). The Cirrus also allows us to see the retinal pigment epithelium (RPE) and all of the retinal layers, making segmentation easier. In the case of central serous retinopathy shown in Figure 1, for example, spectral domain OCT shows some debris in the serous detachment.

3-D Views

Since spectral domain OCT can rapidly analyze the entire macula, a new view of the retina called the 3-D view is possible. In contrast to the 2-D images we see in time domain OCT, 3-D views give depth information and can be segmented.

Figure 1. In this patient who was being treated for central serous retinopathy, spectral domain OCT imaging reveals debris in the serous detachment. This area isn't visible on time domain OCT images.

Figure 2. Time domain OCT shows retinal thickening but doesn't clearly reveal the epiretinal membrane.

Using 3-D layer segmentation, we can examine the RPE or the internal limiting membrane and get a true 3-D retinal thickness map. When I started working with the Cirrus HD-OCT, I was excited to see how well it determined the inner and outer retina boundaries. It's almost spot-on every time. As we learn how to use the 3-D views generated by spectral domain OCT, we'll better understand retinal pathology and, in some cases, even detect disease that would have been missed without this new technology.

For example, when trying to determine if retreatment is necessary in a patient with wet age-related macular degeneration (AMD), we look for cysts and intraretinal swelling. With time domain OCT, we analyze six radial lines that may miss subtle fluid. With spectral domain OCT, we may be able to see very obvious subretinal fluid in multiple areas by paging through the box. In this case, spectral domain OCT reveals key details that we can't see with radial line scans from time domain OCT.

Spectral domain OCT also shows details of the vitreoretinal interface, such as epiretinal membranes. In Figure 2, using time domain OCT, we see that the retina has thickened, but we can't see the epiretinal membrane. The spectral domain 3-D views in Figure 3, however, clearly show the epiretinal membrane.

As we page through the various layers with the Cirrus HD-OCT, we can go deeper and deeper into the eye, and we can see the choroidal vessels, a view we usually don't see using time domain OCT.

Figure 3. A 3-D scan generated by the spectral domain OCT clearly shows the epiretinal membrane.

Greater Precision

If we're participating in a clinical study and following patients with diabetes or AMD, we want to compare how patients are doing from visit to visit. The Cirrus HD-OCT retains the information from a patient's previous visit — such as the position of the head and chin rests and the placement of the scan — and automatically adjusts the hardware. This total patient recall system helps tremendously with ease of use.

For repeatability, the Cirrus offers the repeat scan, which allows the clinician to use an overlay from the previous scan to confirm the precise location for the current scan.

The total patient recall and repeat scan functions provide improved intervisit comparisons that allow us to better evaluate our patients and the effectiveness of their treatments.

Better Outcomes in Retinal Disease

With high speed and high resolution, spectral domain OCT ensures dense coverage of the retina, improves visit-to-visit registration and comparison accuracy, and decreases acquisition times. These advantages will help us diagnose, treat and monitor disease more effectively. My colleagues will discuss more about the clinical applications for this new and exciting diagnostic equipment.

Peter K. Kaiser, M.D., is director of the Digital OCT Reading Center (DOCTR) at the Cole Eye Institute, Cleveland Clinic Foundation.