Best Practices in AMD Management

Learn why Cirrus HD-OCT is the ideal instrument for wet — and dry — AMD.

By Philip J. Rosenfeld, MD, PhD

For retinal specialists who see a high volume of AMD patients daily, the Stratus OCT (Carl Zeiss Meditec, Dublin, Calif.) is a useful instrument. For wet AMD in particular, the Stratus allows us to see everything we need to see to provide quality care. However, I have made the switch to spectral domain OCT because it is faster and easier, and it gives me greater confidence in my wet AMD diagnostic and management decisions. Furthermore, when it comes to dry AMD, SD-OCT is revolutionizing the way we manage our patients.

In this article, I explain how I use SD-OCT in practice, specifically the Cirrus HD-OCT.

Faster Scanning Produces More Data

The raster scanning pattern utilized by Cirrus HD-OCT has changed the way we generate images. In less than 2 seconds, 40,000 A-scans (200 A-scans x 200 B-scans) are obtained. In less than 3 seconds, 65,536 A-scans (512 A-scans x 128 B-scans) are obtained. The result is an astounding set of information covering a 6-mm x 6-mm cube.

The information, of course, is represented for us on a printout. The printout for the dense macular cube scan (512 x 128), for example, includes the horizontal and vertical B-scans. It also includes 3-D segmentation maps of the ILM and the RPE (Figure 1). The segmentation techniques also make it possible to reconstruct 2-D and 3-D thickness maps (Figure 2).

Figure 1. Segmentation techniques make it possible to view retinal layers individually, such as the internal limiting membrane and retinal pigment epithelium.

Figure 2. From the large amount of data obtained, the Cirrus HD-OCT reconstructs 2-D and 3-D thickness maps.

Macular Change Analysis in Clinical Practice

The following cases from my practice demonstrate how the newest Cirrus HD-OCT imagecapture and analysis algorithms improve clinical decision-making.

Case 1

In the first case, the patient presented with a hemorrhagic pigment epithelial detachment (PED) and visual acuity of 20/30⁺². The diagnosis was relatively straightforward. Fluorescein angiography showed a well-defined PED and blockage from the blood. Indocyanine green angiography showed neovascular tissue creeping up the side of the PED. When we scanned the patient with Cirrus HD-OCT, the 2-D and 3-D thickness maps revealed a ring of fluid, and the segmented view of the RPE revealed the PED (Figure 3). The Stratus OCT would have shown us the same features, just not in such elegant, three-dimensional detail. Using the Cirrus HD-OCT Multi-Slice Report, which displays a sampling of B-scans (Figure 4), we confirmed the presence of subretinal fluid. We treated the patient with an intravitreal injection of an anti-VEGF agent, bevacizumab (Avastin, Genentech).

Figure 3. Two- and three-dimensional thickness maps reveal a ring of fluid in this eye with a hemorrhagic pigment epithelial detachment (PED). The PED is clearly visible on the RPE segment map.

Figure 4. Confirmation of subretinal fluid using the Cirrus HD-OCT Multi-Slice Report.

One month later, when the patient returned for scheduled follow-up, the thickness maps indicated the amount of macular fluid had decreased. The elevation was less and the horizontal and vertical B-scans showed the PED was smaller. However, some fluid was still present. From this point, we treated the patient with three additional injections of bevacizumab at scheduled intervals.

At the next follow-up visit, 1.5 months after the fourth injection (Figure 5), little to no fluid was visible on the thickness map. However, on the highdefinition crosshair image, which provides enhanced resolution at the center of the scan, there appeared to be a very small amount of fluid. Also the PED was still present. Based on what we were seeing, combined with the patient's desire to delay treatment, we decided to observe rather than treat the patient.

Figure 5. While the amount of subretinal fluid in the patient's eye was nearly gone 1.5 months after a fourth injection of bevacizumab (Avastin, Genentech), the high-definition crosshair image indicated the possibility of a small amount remaining. The PED was still present. Based on this information, it was decided the patient would be observed rather than treated again.

When the patient returned to the office 3 months after the fourth treatment with bevacizumab, it was immediately clear from the thickness map derived from the 200x200 Macular Cube Scan that more subretinal fluid had returned (Figure 6). Not only were we able to see the fluid on the map, but we were also able to confirm the status of the retina using the Cirrus HD-OCT's Macular Change Analysis feature.

Figure 6. When the patient returned to the office 3 months after the fourth treatment with bevacizumab, it was immediately clear from the thickness map derived from the 200 x 200 Macular Cube Scan that macular fluid had returned.

The Macular Change Analysis algorithm allows us to automatically compare results from one patient visit to the next. It takes two HD-OCT fundus images and aligns them. The thickness map from one visit is subtracted from the thickness map from another visit, which results in a perfectly registered difference map. As we move the horizontal B-scan on the first image, the B-scan moves accordingly on the second image. This permits the correlation of registered B-scans from one visit to the next. The printout shows the thickness maps from the two visits, along with the difference in thickness.

In this patient, the brown color on the difference map indicated for us how the status of the retina had changed — fluid had returned since the last visit (Figure 7). The patient was treated subsequently with two more injections of bevacizumab. Over time, all of the fluid resolved and the PED slowly disappeared (Figure 8).

Figure 7. The Macular Change Analysis printout shows the thickness maps from two visits along with the difference in thickness.

Figure 8. As the patient was followed over time, Cirrus HD-OCT high-resolution scans and Macular Change Analysis showed resolution of subretinal fluid and the PED.

Case 2

This second case (Figures 9-12) also illustrates how the Macular Change Analysis feature improves our ability to monitor pathology and helps us more easily determine when treatment is necessary.

Figure 9-12. Cirrus HD-OCT images from a patient with AMD over a 4-month period. The Macular Change Analysis function allowed precise monitoring of a gradual increase in subretinal fluid that eventually required anti-VEGF treatment. One month after treatment, the map comparing the current and previous visits showed a decrease in fluid, which indicated the therapy was effective.

The patient is an 82-year-old woman with AMD. At one particular visit, a small cyst was visible on the Cirrus HD-OCT scan. To determine if fluid had accumulated since her previous visit, we checked the Macular Change Analysis. It showed no change, so we decided to observe. When the patient returned 8 weeks later, Macular Change Analysis detected a slight increase in the amount of macular fluid. Again, we elected not to treat but to have the patient come back in approximately 2 weeks. During the subsequent visit, the macular change difference map clearly identified a large area of increased fluid. We treated with an intravitreal injection of ranibizumab (Lucentis, Genentech). When the patient returned a month later for follow-up, the difference map indicated the fluid had diminished.

SD-OCT Fundus Image and Dry AMD

In the previous section, I mentioned the OCT fundus image produced by the Cirrus HD-OCT instrument and its role in the Macular Change Analysis function. The OCT fundus image, which was not available with Stratus OCT, is also particularly useful for following patients with dry AMD.

To understand why, remember that the A-scans captured with Cirrus HD-OCT make up B-scans, and the B-scans make up the 3-D dataset, or SD-OCT cube. The OCT fundus image is a projection of all of the summed reflectivity of the compiled B-scans projected as a virtual 2-D fundus image. This virtual fundus image is depicted on the top surface of the cube. It provides a great representation of the macula and it can be aligned with any other fundus image to achieve point-to-point correlation (Figure 13).

Figure 13. The SD-OCT fundus image is a projection of all the summed reflectivity and visualized on the top surface of the 6-mm x 6-mm data cube. Because it provides a great representation of the macula and can be aligned point-by-point with any other fundus image, its usefulness extends to following dry AMD.

The OCT fundus image provides excellent visualization of geographic atrophy (GA) because the 840nm light from the HD-OCT does not penetrate well into the choroid where the RPE is intact. In areas where the RPE is absent, the light has greater penetration into the choroid. Therefore, wherever the RPE is intact, there is less reflectivity, a darker image. Wherever the RPE is not intact, as in GA, there is more reflectivity, a brighter image.

We have published a study in a small number of patients showing that SD-OCT can identify and quantitate areas of GA, and the size and shape of these areas correlate well to the areas of GA seen using fundus autofluorescence (FAF).¹ We also showed that manual identification of GA using the SD-OCT fundus image is reproducible and could be used as a practical method to quantify the presence and progression of GA in a clinical trial. To measure GA growth, we superimpose the color fundus photograph, the FAF image and the SD-OCT fundus image. By hand-drawing the outline of the GA at different follow-up time points, we can calculate the enlargement area and rate (Figures 14-16). Going forward, this is how we will evaluate potential new therapies for dry AMD — by testing whether treatments slow or stop the enlargement rate of GA.

Figure 14.

Figure 15.

Figure 16. In Figures 14-16, Cirrus HD-OCT can be used to quantify the presence and progression of geographic atrophy

A new algorithm is under development for the Cirrus HD-OCT that will allow automatic, rather than manual, quantification of GA. This will be an exciting advance for reading centers, and it will also allow those of us in clinical practice to follow our patients much more closely and better explain why they are losing vision.

Measuring Drusen Area and Volume

Also by the end of this year, we will be able to use the Cirrus HD-OCT to measure drusen reliably. Because the instrument can register fundus images and segment the ILM and RPE, we can take advantage of the direct correlation of drusen and retinal thinning.

The new algorithm creates a difference map that quantitates drusen by subtracting the "interpolated normal RPE" from the "real RPE" segmentation map that contains the drusen. This gives us the ability to measure the area and volume of drusen.

In addition, if we think of serous PEDs simply as RPE deformation that are similar to drusen, then we can see why the same algorithm that quantitates drusen in dry AMD will also provide reliable and reproducible measurements of PEDs in wet AMD (Figures 17-18).

Figure 17-18. A soon-to-be-released algorithm for the Cirrus HD-OCT will allow automatic area and volume measurements of drusen and pigment epithelial detachments. In this case, even though the area of the PED does not appear to be changing significantly, treatment has decreased the volume of the PED by nearly 80%.

Moving Forward to Meet the Future

The Cirrus HD-OCT software advances discussed in this article enhance our ability to manage patients with wet AMD. At the same time, they revolutionize how we can manage our patients with dry AMD. When a successful therapy for dry AMD becomes available, we will have an ideal combination of imaging and therapy.

---

Dr. Rosenfeld is a Professor of Ophthalmology at the University of Miami Miller School of Medicine's Bascom Palmer Eye Institute.

REFERENCE

1. Lujan BJ, Rosenfeld PJ, Gregori G, et al. Spectral domain optical coherence tomographic imaging of geographic atrophy. Ophthalmic Surg Lasers Imaging 2009;40:96-101.