AMD DIAGNOSTICS
Next-Generation Optical Coherence Tomography
Spectral domain technology promises improved image resolution and added capabilities.

Figure 1. Spectral domain OCT mapping of retinal pigment epithelium drusen. The thickness map clearly shows that the retina is thinner over the drusen.

According to Philip J. Rosenfeld, MD, PhD, new optical coherence tomography (OCT) instruments, which will likely be commercially available within the next year, will offer enhanced capabilities. At the 2006 Retinal Physician Symposium (May 31-June 3, Atlantis, Paradise Island, Bahamas), he introduced the Miami Macula Mapper, an enhanced version of the technology, which has been developed by Carmen A. Puliafito, MD, MBA, Robert Knighton, PhD, Shuliang Jiao, PhD, and Giovanni Gregori, PhD, of the Bascom Palmer Eye Institute's OCT Laboratory.

The Miami Macula Mapper uses a spectral domain detection technique, rather than the time domain technique used by currently available OCT instruments. "Spectral domain is next-generation OCT," Dr. Rosenfeld said.

Time domain OCT makes use of one detector, and each A-scan requires the movement of several mechanical parts. Rather than a single detector, spectral domain OCT uses a spectrometer with up to 2000 detector elements and has no moving parts, which means information is collected much faster. Spectral domain OCT's imaging speed is 29,000 A-scans per second compared with time domain OCT's 400 A-scans per second.

Faster acquisition speed produces more sensitive, higher-resolution images, including 3-D images, which significantly improves visualization of retinal structures. The Miami Macula Mapper is capable of 4-μm resolution. "You can break up the scans into as few or as many as you want," Dr. Rosenfeld said. "There is much more flexibility in the resolution that can be obtained." Furthermore, retinal layers can be segmented and represented by 3-D thickness maps (Figure 1).

Spectral domain OCT can also generate a fundus image similar to that acquired with a scanning laser ophthalmoscope, and the spatial location of each OCT sectional image is registered on the OCT fundus image automatically (Figure 2). "Spectral domain has point-to-point registration in reference to surface landmarks," Dr. Rosenfeld explained. "It allows users to precisely place where the pathology is in the different layers of the retina based on the vascular landmarks of the fundus image."

The sophisticated algorithms used in spectral domain OCT permit more elegant, clinically useful analyses, Dr. Rosenfeld said. He also clarified that spectral domain OCT is known as "high-resolution," which is different from "ultra high-resolution." The

Figure 2. Spectral domain OCT can generate a fundus image, and the spatial location of each OCT sectional image is registered on the fundus image automatically.

ultra high resolution OCT technology, which is currently being used as a research device at New England Eye Center in Boston and a site in Vienna, will not be commercially available.

Figure 1. In this patient, one anti-VEGF treatment reduced diffuse edema, but did not eliminate sub-RPE fluid. Two subsequent treatments eliminated all subretinal, intraretinal and diffuse edema. Visual acuity improved from 20/70 to 20/40.

Qualitative OCT in Wet AMD
Studies shed light on using OCT to track the efficacy of anti-VEGF therapies.

With the availability of anti-VEGF therapies marking the beginning of a new era in the treatment of neovascular age-related macular degeneration (AMD), the role of optical coherence tomography (OCT) becomes even more vital. According to David M. Brown, MD, "As the PrONTO study has shown, OCT is absolutely essential for assessing response to an anti-VEGF agent and for determining when to retreat."

At the 2006 Retinal Physician Symposium (May 31-June 3, Atlantis, Paradise Island, Bahamas), Dr. Brown emphasized the importance of using the qualitative rather than the quantitative data provided by OCT. He cited a recently published paper by Sadda, et al.¹,which indicated that retinal thickness measurement errors can and do occur frequently with the current Stratus OCT (Carl Zeiss Meditec) segmentation and analysis algorithms in patients with neovascular AMD. Errors are inherent, Dr. Brown explained, because scans for computing center-point thickness require patient fixation, retinal boundaries need to be determined, and subretinal pigment epithelium (sub-RPE) fluid is not taken into account. "In AMD, determining retinal boundaries is difficult because of schisis channels and the different layers," he said. In addition, eccentric lesions can interfere with central subfield measurements.

Figure 2. After this patient received four anti-VEGF treatments, subretinal fluid, diffuse edema, intraretinal cysts and sub-RPE fluid remained. A subsequent treatment with a different anti-VEGF agent did not eliminate the intraretinal cysts. The patient was treated again with the first agent, to avoid further elevation of intraocular pressure, but fluid returned. Treatment with a third anti-VEGF agent led to the desired anatomic response.

"With qualitative OCT, what you see is what you get," Dr. Brown said. Therefore, physicians monitoring the effects of anti-VEGF treatments with OCT should examine each high-resolution scan for signs of fluid reaccumulation, as was done in the PrONTO study. When fluid returns, it does so in the form of diffuse edema, intraretinal cysts, subretinal edema, or sub-RPE fluid, and it means the disease is "DISSing" — disrespecting — the treatment, Dr. Brown said. "Qualitative OCT, 'DISS,' equals leakage, which equals disease activity. If AMD is DISSing your treatment, treat more often or change agents" (Figures 1 and 2).

As the 1-year results from the PIER and PrONTO studies showed, no treatment or treatment schedule has the same effect in every patient, Dr. Brown said. "There-fore it is essential to determine anatomic response to treatment with qualitative OCT, tailor retreatment to patient response, and change agents when necessary."

Speaking with Retinal Physician after the symposium, Dr. Brown said quantitative OCT values are useful for monitoring diffuse diabetic macular edema, but he never uses them to guide treatment for AMD patients. After he uses fluorescein angiography to make the initial diagnosis of neovascular AMD, he follows patients with OCT.

"Just like a macula-off retinal detachment has to be anatomically corrected before visual acuity can be maximized, I believe in AMD the best visual acuity gains occur when the retina is entirely deturgesced," he said. "I treat with anti-VEGF agents monthly until all signs of fluid are gone; no diffuse edema, no intraretinal cysts, no subretinal fluid and no sub-RPE fluid. I then follow patients with the retinal thickness map protocol on the OCT3 (six high-resolution cuts 30° apart). I see them every 6 weeks and retreat at the first sign of any recurrent leakage."

REFERENCE

1. Sadda SR, Wu Z, Walsh AC, et al. Errors in retinal thickness measurements obtained by optical coherence tomography. Ophthalmol. 2006;113: 285–293.

Simple Scale Scores Probability of AMD Progression
In-office assessment provides practical approach to patient prognosis.

A simple scale to estimate the risk of AMD progression based on the presence of large drusen and pigment abnormalities may improve early detection of choroidal neovascularization (CNV), according to Susan B. Bressler, MD, the Julia G. Levy, PhD professor of ophthalmology, Johns Hopkins University School of Medicine, The Wilmer Eye Institute. Dr. Bressler spoke on behalf of the AREDS Research Group at the May 2006 Retinal Physician Symposium.

The development of the in-depth AMD research severity scale by the AREDS Research Group led to the formulation of the simplified AMD severity scale.¹ The format of the simplified scale was suggested by three observations that were recognized during the construct of the research scale:

1. A strong association between drusen area and largest drusen size

2. A low frequency of retinal pigment epithelium depigmentation and/or geographic atrophy in the absence of increased pigment

3. Bilateral large drusen is a stronger risk factor than unilateral large drusen.

The simplified scale should enable clinicians to differentiate AMD patients into risk profiles, making it practical to target those who are most likely to benefit from preventive measures and more apt to progress as well as benefit from intensive monitoring, Dr. Bressler said.

"We made the assumption that a clinician would be more likely to effectively assess maximum drusen size in the posterior pole as compared to maximum area occupied by drusen," Dr. Bressler said. "Fur-
thermore, we thought clinicians could [more easily] sort patients by those who had large drusen and those who did not."

For the purposes of this scale, Dr. Bressler noted, large drusen are defined as a drusen within two disc diameters of the foveal center, with a diameter of at least 125 microns. "Conveniently, this is equal to the average width of a retinal vein at the disc margin," Dr. Bressler said.

In addition to drusen size, the AMD severity scale requires the clinician to assess the presence or absence of definite pigmentary changes. "This may consist of either focal increases in pigment or areas of depigmentation within one disc diameter of the fovea or areas of non-central geographic atrophy within two disc diameters of the fovea," Dr. Bressler said.

SCORING THE RISK

The severity scale provides a summary score from 0 to 4 that differentiates a patient's risk of developing advanced AMD over 5 years, from a low of <1% to as high as 50%. For example, an individual without large drusen and without pigment abnormalities in both eyes would have zero risk factors; whereas, an individual with large drusen and pigment abnormalities in both eyes would have four risk factors.

The scores correspond to a marked differential in the risk of developing advanced AMD in one or both eyes. Zero risk factors imply a negligible risk for advanced AMD. One risk factor implies a small 5-year risk of 3%.

"There is at least a doubling of risk for each successive step with two risk factors corresponding to a 12% risk; three risk factors corresponding to a 25% risk; and four risk factors corresponding to a 50% risk," Dr. Bressler said. "If you are examining a patient who has advanced AMD in one eye, and you want to assess his or her risk of developing advanced AMD in the remaining eye, you would score the eye with advanced AMD as a 2, and then continue on to the eye at risk and assess it for large drusen and pigment changes," Dr. Bressler said. "If the eye at risk has neither, then the patient's total score is 2, with roughly a 12% 5-year rate of progression in the fellow eye. Whereas if the fellow eye has large drusen and pigment changes, then the fellow eye is at the high risk rate of 50%."

An individual who does not have large drusen in either eye, but has bilateral extensive intermediate size drusen, would be assigned a score of one-half for each eye, or a total of 1 for the patient. If neither eye has a pigment change, then the total patient score is 1, which implies a 3% risk that one or both eyes will develop advanced AMD within 5 years.

GOAL: CLINICAL UTILITY

"Our objective was to develop a scale that could be useful in the clinic, and we hope this has been accomplished by concentrating on these two characteristics that can be assessed on routine ophthalmoscopy or review of fundus photographs," Dr. Bressler said. "With this scale, you may be able to determine who you want to follow more often to potentially improve the early detection of CNV.

REFERENCES:

1. AREDS Report #18: Arch Ophthalmol. 2005;123:1570-1574.