Practical Applications of SD-OCT Imaging in Dry AMD

This technology has revolutionized the care of wet AMD. Can it help in patients with dry AMD?

BRANDON J. LUJAN, MD

The profound benefits of optical coherence tomography imaging in age-related macular degeneration first became apparent as a clinical tool coinciding with the rise of anti-vascular endothelial growth factor therapies in treating exudative or wet AMD. Subretinal and intraretinal fluid could be visualized on OCT that was not clinically evident or fully appreciated by fluorescein angiography. The modulation of this fluid through treatment fostered an increased appreciation of the pathogenesis of wet AMD, and it has served as a decision point in treatment paradigms of wet AMD for retinal physicians worldwide.¹

Conversely, nonexudative or dry AMD does not directly benefit from OCT in the same way. This is unfortunate, because dry AMD is much more common than the wet form and indeed continues to progress even in the face of treated wet AMD. Because of the current lack of a therapeutic intervention, there is no clinical decision point for which OCT is employed to direct. The standard of care for this disease — typified by drusen, retinal pigment epithelium disturbances and geographic atrophy — in its intermediate or advanced form continues to be the use of AREDS formula vitamins, given their beneficial effect.² Other than this and smoking cessation, there is currently no other intervention.

Despite this lack of OCT's role as a current clinical decision maker, in vivo imaging has a strong scientific role. With the advent of spectral-domain OCT, an increasingly profound understanding of the pathogenesis of dry AMD is beginning to emerge through clinical research. Having good data regarding the natural history of AMD using SD-OCT is critical, as several different clinical trials are now under way to attempt to combat dry AMD on several fronts. These strategies include drug targets intended to preserve photoreceptors and the RPE, drugs to diminish the effects of oxidative stress and nutrient depletion, and immunomodulatory agents.³ The use of SD-OCT to monitor treatment response and assess disease progression will be critical to defining success in these clinical trials.

SD-OCT MECHANICS

The essential principles all OCT systems rely upon are reflectivity and interferometry. The former dictates that light will be scattered back from the optical interfaces it encounters. In this respect, OCT is the light analogue of ultrasound or sonar. OCT differs from these, however, in that the speed of light necessitates the use of interferometry to indirectly measure coherent light returning from two arms, each having the same optical path length. In all retinal OCT systems, one of those arms is the retina and the other is a reference mirror, and the intensity of coherently interfering light emerging from where these paths cross can be measured.⁴

The original models of OCT, culminating commercially in Stratus OCT (Carl Zeiss Meditec, Dublin, CA), is called a time-domain system because the reference mirror moves in time and interference is detected through positions of equal optical path length to the retina. The speed of the system is dictated by the need to move this reference mirror during the acquisition of each axial scan (A-scan). Despite this requirement, Stratus OCT is capable of acquiring 400 A-scans per second, which permits a cross-sectional sampling through the retina (B-scan).

Spectral-domain OCT, also known as Fourier-domain OCT, revolutionized OCT technology by recognizing that a moving reference mirror was extraneous and that in-depth information could be obtained from analysis of interference fringes off a fixed reference mirror and signal processing utilizing Fourier transformation.⁵ By eliminating the need for a moving reference mirror and by using broadband light sources, both the speed and axial resolution of SD-OCT systems are markedly improved. Commercial systems acquire between 20,000 and 40,000 A-scans per second. This increased speed permits at least two advances over time domain OCT: volumetric imaging and frame averaging.

Volumetric imaging is accomplished by performing a dense raster scan over the macula. The spacing between adjacent A-scans in both the horizontal and vertical direction can approach the lateral resolution limits imposed by the eye's optical aberrations. The volumetric data obtained can be used in several ways. One application available to volumetric OCT is the ability to generate an OCT fundus image, also referred to as a summed voxel projection, which is obtained by adding the pixel values along each A-scan and normalizing the results to be viewed en face.⁶ Dense raster scanning also allows segmentations of retinal features such that identification of a specific layer on each B-scan can be visualized as a continuous plane when aggregated.

The speed of SD-OCT also permits frame averaging by simply utilizing the speed of the instrument or by using an eye tracker. Frame-averaging strategies acquire B-scans only slightly offset from each other vertically. The averaging of these adjacent B-scans increases the true retinal signal by dampening the speckle noise intrinsic to interferometry. Both strategies, cross-sectional and volumetric imaging, have been employed to address the pathogenesis of dry AMD.

CURRENT SD-OCT FINDINGS IN DRY AMD

Drusen and Pigment Migration

As measured by fundus photography in the AREDS trial, increased size and distribution of drusen, along with pigmentary changes, are predictive for the development of the advanced forms of AMD, central GA and progression to wet AMD.⁷ More recently, a retrospective analysis of patients enrolled in AREDS who were determined to have drusenoid pigment epithelial detachments at baseline went on to have poor vision as a result of pigmentary changes and calcified drusen, GA, or progression to wet AMD.⁸ While, clinically and photographically, drusen vary in their size, borders and distribution, only with SD-OCT can the reflective cross-sectional heterogeneity of drusen be appreciated. An important pilot study of drusen microstructure demonstrated variability in reflectivity within the core of individual druse, as well as identified associated features such as pigment migration above drusen.⁹ The classification of drusen subtypes was found to be reproducible among trained graders. Even in adjacent drusen (Figure 1), differences in reflectivity are observable, as well as their impingement upon the outer nuclear layer. Indeed, one study found that, in a cohort of patients with dry AMD, 97% of the drusen analyzed had associated thinning of this layer where the photoreceptor nuclei reside.¹⁰

Figure 1. Bioptigen frame-averaged image of drusen, courtesy of Joe Carroll, PhD, Medical College of Wisconsin. Note differences in reflectivity in druse above asterisks from surrounding drusen, as well as the decrease in the outer nuclear layer immediately above the drusen.

This ability to “look inside” drusen and directly study their effects on the surrounding retina and RPE in vivo should drive the acquisition and analysis of these data. It is tantalizing to believe that this may ultimately be more sensitive than fundus photography alone. Identification of different subtypes of drusen based on their SD-OCT features is the important first step in determining which categories of drusen predict worse outcomes. Only long-term, consecutive SD-OCT imaging can educate us as to which of these phenotypical SD-OCT patterns convey higher risks of AMD progression.

Geographic Atrophy

Unlike clinical trials of disease processes mediated by choroidal neovascularization and retinal edema, current and future clinical trials of treatments directed at dry AMD cannot simply use visual acuity as a clinical endpoint. Because the advanced form of dry AMD is a slow-moving process and the fovea may be spared until late in the disease course, acuity can remain stable while the area of atrophic retina may continue to grow. Consequently, the progression of GA has gained acceptance as an endpoint measurement in clinical trials targeting dry AMD.¹¹

Spectral-domain OCT is able to visualize RPE atrophy in cross-section by the absence of melanin, the RPE constituent responsible for its high reflectivity. When absent, there is increased deep reflectivity stemming from the unimpeded penetration of light allowed when RPE melanin is absent. The cross-sectional morphology of the retina immediately adjacent to large areas of atrophy has been studied.¹² The consistent feature immediately adjacent atrophy is the collapse of the outer retinal layers. Similar to imaging of drusen, a wide range of morphological alterations were identified at the level of the RPE and outer retina. The reflectivity of the retina adjacent to geographic atrophy may be a snapshot into the agonal moments of the RPE life cycle in the face of dry AMD; however, longitudinal studies of the same retinas over time are required before a pathological mechanism can truly be postulated.

In addition to cross-sectional images performed through areas of geographic atrophy, volumetric SD-OCT protocols employ a dense raster scan and are capable of forming an OCT fundus image, providing an en face image capable of identifying and quantifying the area of GA.¹³ This is a direct consequence of deep hyper-reflectivity from the underlying choroid when RPE melanin is absent. This increased deep hyper-reflectivity directly makes the area of GA bright on the OCT fundus image because the A-scans within the region of atrophy have a higher relative value compared to regions outside the atrophy, where the choroid does not backscatter as much signal because of an intact RPE (Figure 2).

Figure 2. Registered Cirrus (Carl Zeiss Meditec) OCT fundus image to fundus photograph demonstrating the appearance of geographic atrophy, based on the deep choroidal penetration visible on B-scan.

Volumetric Drusen Analysis

In addition to the merits of identifying GA area by manipulating volumetric data, SD-OCT is unique in its ability to quantitate drusen volume. While fundus photography has been the gold standard for determining drusen burden on the macula, it can evaluate only the area of drusen present. Registration of OCT cross-sectional images has demonstrated that what may appear to be drusen on examination and fundus photography may not be volume-occupying lesions, as expected, but rather an early atrophic patch.

Correlation between AREDS-style color photographic grading and manually segmented SD-OCT drusen has been reported.¹⁴ This pilot study found that there was a correlation between drusen volume and the photographic appearance of drusen, though this study was limited because of sparse sampling of SD-OCT data and the fact that drusen volume was inferred from drusen area and not directly measured. Future investigations of correlation between drusen appearance on fundus photographs and SD-OCT would benefit from a dense sampling, which would confer a direct volume estimate.

While manual segmentation is currently the gold standard for drusen volume calculations, the development of a reliable automated algorithm is critical to use of SD-OCT data in future clinical trials. Not only is manual grading time consuming, but it is also subject to variability in which the grader places segmentation points within the RPE and Bruch's membrane. Consistent offsets between graders of only a single pixel could yield a significant intergrader discrepancy over an entire macular volume.

The holy grail of SD-OCT imaging of dry AMD imaging is a robust, reliable algorithm for the accurate determination of drusen area and volume, as well as the quantification of GA area. This would provide a single snapshot of the macula (Figure 3) that would encompass metrics for the defining features of dry AMD and that could be followed over time. However, an automated algorithm for drusen areas and volumes has yet to be validated or approved by the FDA. The first software algorithms to achieve this will certainly have an important role to play in current and future dry AMD studies.

Figure 3. Fundus photograph, overlayed Cirrus (Carl Zeiss Meditec) OCT fundus image demonstrating GA, and hybrid image demonstrating quantification of the encircled GA area, as well as drusen volume and area based on an algorithm, developed by Giovanni Gregori, PhD, of Bascom Palmer Eye Institute.

FUTURE PROGRESS OF SD-OCT IN DRY AMD

The future of spectral-domain OCT imaging technology is bright. Advances in light-source technology that produce broader bandwidths will likely allow higher-resolution imaging. Longer-wavelength systems that better visualize the choroid will also likely come to fruition. Advances in light sources and photon-capturing technologies will similarly allow systems to operate at even higher speeds.

Spectral-domain OCT technology, however, has already outstripped the pace of the scientific discoveries of in vivo dry AMD. While many important studies have been carried out, the true benefits of SD-OCT monitoring of dry AMD are admittedly unclear without longitudinal analysis. Despite this fact, our profession's collective intuition is that the high-resolution, cross-sectional and volumetric imaging that SD-OCT provides will be critical to a deeper understanding of disease progression. This will only be enhanced by the further development of genetic and proteomic biomarkers in AMD research that correlate to phenotypic expression of disease.

Longitudinal SD-OCT data that will be acquired by clinical studies directed at treating dry AMD, as well as data acquired from contralateral dry AMD eyes in clinical trials of wet AMD, will be incredibly valuable in improving our understanding of AMD pathology and advancing our ability to treat this devastating disease. RP

REFERENCES

1. Rosenfeld PJ, Rich RM, Lalwani GA. Ranibizumab: Phase III clinical trial results. Ophthalmol Clin North Am. 2006;19:361-372.
2. Age-related eye disease study research group. A randomized, placebo-controlled, clinical trial of high-dose supplementation with vitamins C and E, beta carotene, and zinc for age-related macular degeneration and vision loss: AREDS report no.8. Arch Ophthalmol. 2001;119:1417-1436. Erratum in: Arch Ophthalmol. 2008;126:1251.
3. Rosenfeld PJ, Legarreta JE. Preclinical and phase 1 drugs in development for dry AMD. Retinal Physician. 2009;6(9):27-32.
4. Huang D, Swanson EA, Lin CP. Optical coherence tomography. Science. 1991;254:1178-1181.
5. Wojtkowski R, Leitgeb A, Kowalczyk T, et al. In vivo human retinal imaging by Fourier domain optical coherence tomography. J Biomed Opt. 2002;7:457-463.
6. Jiao S, Knighton R, Huang X, Gregori G, Puliafito C. Simultaneous acquisition of sectional and fundus ophthalmic images with spectral-domain optical coherence tomography. Opt Express. 2005;13:444-452.
7. Ferris Fl, Davis MD, Clemons TE, et al. A simplified severity scale for age-related macular degeneration: AREDS Report No 18. Arch Ophthalmol. 2005;123:1570-1574.
8. Cukras C, Agrón E, Klein ML, et al; Age-Related Eye Disease Study Research Group. Natural history of drusenoid pigment epithelial detachment in age-related macular degeneration: Age-Related Eye Disease Study Report No. 28. Ophthalmology. 2010;117:489-499.
9. Khanifar AA, Koreishi AF, Izatt JA, Toth CA. Drusen ultrastructure imaging with spectral domain optical coherence tomography in age-related macular degeneration. Ophthalmology. 2008;115:1883-1890.
10. Schuman SG, Koreishi AF, Farsiu S, et al. Photoreceptor layer thinning over drusen in eyes with age-related macular degeneration imaged in vivo with spectral domain optical coherence tomography. Ophthalmology. 2009;116:488-496.
11. Csaky KG, Richman EA, Ferris FL 3rd. Report from the NEI/FDA Ophthalmic Clinical Trial Design and Endpoints Symposium. Invest Ophthalmol Vis Sci. 2008;49:479-89.
12. Fleckenstein M, Issa PC, Helb H-M, et al. High-resolution spectral domain-OCT imaging in geographic atrophy associated with age-related macular degeneration. Invest Ophthalmol Vis Sci. 2008;49:4137-4144.
13. 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.
14. Freeman SR, Kozak I, Cheng L, et al. Optical coherence-tomography raster scanning and manual segmentation in determining drusen volume in age-related macular degeneration. Retina. 2009;10:1-5.