Fundus Autofluorescence in Retinal Disease: A Review and Perspectives

Part one of two parts

MADELINE YUNG, MD • DAVID SARRAF, MD

Due to its noninvasive nature, ease of use, and sensitivity to detect various retinal pathologies, fundus autofluorescence (FAF) imaging has gained a solid foothold in clinical practice, and the scope of its applications continues to grow. This article describes the molecular origins of autofluorescence, compares common FAF imaging systems, and highlights several retinal diseases with characteristic FAF findings (Table 1).

MOLECULAR ORIGINS OF AUTOFLUORESCENCE

Fluorophores are compounds with the capacity to absorb light at specified wavelengths (excitation spectrum) and then to release light at different wavelengths (emission spectrums). FAF imaging systems illuminate the retina with an excitation beam and then collect the emitted autofluorescence from ocular fluorophores to form a brightness map.¹

Ocular Fluorophores

Classically, FAF uses a short-wavelength, blue-light excitation beam (SW-FAF) to construct a topographic map of lipofuscin, the predominant ocular fluorophore with a peak absorption at 470 nm and emission at 600-610 nm.²

Lipofuscin is a heterogeneous mixture formed from the degradation of photoreceptor outer segments, and it accumulates in retinal pigment epithelium lysosomes with age.^1,3 The distribution of lipofuscin reflects the overall health of the RPE and is altered in various retinal diseases. In healthy eyes, the lipofuscin concentration is greatest in the posterior pole, is limited in the fovea, and decreases toward the periphery.^1,4

Other ocular fluorophores include vitelliform material and optic disc drusen. Named after their orange yolk-like appearance, vitelliform lesions represent the extracellular deposition of shed outer segment debris in the subretinal space.^5-7 Optic disc drusen derive from deposits of extracellular mitochondria in a filamentous protein matrix, and they produce hyperautofluorescence if superficial.⁸

Other Factors That Influence Autofluorescence

Compounds such as melanin, rhodopsin, and macular pigments can absorb light, prevent the excitation beam from reaching ocular fluorophores, and weaken the autofluorescent signal.^3,9,10 Located in RPE cells and uveal melanocytes, melanin absorbs short-wavelength light to protect the retina from light-induced damage.^11-13 Notably, melanin does autofluoresce at longer wavelengths, with peak excitation at 787 nm, and it is the primary fluorophore for near-infrared FAF (NI-FAF).¹⁴

Rhodopsin, a visual pigment found in the rod photoreceptor outer segments, initially attenuates autofluorescence but undergoes photoisomerization with continued light exposure, resulting in the bleaching effect that leads to an increase of background autofluorescence of as much as 30% over time (Figure 1).^15,16 The macular pigments lutein, zeaxanthin, and meso-zeaxanthin also filter blue light and protect the retina from photo-oxidative damage.^17,18

Figure 1. Bleaching effect due to rhodopsin. Immediately after obtaining a 30 degree FAF (a), the 50 degree FAF (b) shows a discrete area of relative hyper-autofluorescence corresponding to an area of bleached photoreceptors resulting from the first image acquisition.
COURTESY OF BIOMED CENTRAL

Autofluorescence from the cornea and lens can interfere with and degrade image quality. The cornea is excited at 365-480 nm and emits at 620 nm,¹⁹ while the lens is excited at 420-430 nm and emits at 520 nm.²⁰ Scattered light from these structures can falsely increase the autofluorescent signal, called pseudoautofluorescence.²¹

INTERPRETATION OF FAF FINDINGS

Taken together, the cumulative influences of ocular fluorophores, absorptive compounds, and autofluorescent structures anterior to the retina all contribute to form the FAF image.²² FAF abnormalities are classified as hyperautofluorescent (increased signal), hypoautofluorescent (decreased signal), or isoautofluorescent.

Hyperautofluorescence can result from the accumulation of lipofuscin or other ocular fluorophores due to RPE dysfunction, the bleaching effect, or a window defect from the degeneration of overlying photoreceptors (Figure 2A). Hypoautofluorescence can result from the absence of RPE secondary to atrophy or tearing and blockage or shadowing of the RPE by overlying fluid, structures, or vitreous opacities (Figure 2B). Abnormalities that do not affect the RPE or photoreceptor layers may remain isoautofluorescent.

Figure 2. Causes of hyperautofluorescence (a). Causes of hypo-autofluorescence (b).
COURTESY OF BIOMED CENTRAL

FAF IMAGING SYSTEMS

Fundus Cameras

Fundus cameras, such as those produced by Topcon (Oakland, NJ), Carl Zeiss Meditec (Dublin, CA), and Canon (Melville, NY), utilize single flash (flood light) illumination to produce color autofluorescence images (Figure 3A).²³ The single flash of light is more comfortable for patients and renders fundus cameras inherently resistant to motion artifacts, but in turn it yields lower resolution images.

Figure 3. Comparison of common imaging systems available for fundus autofluorescence. Images of normal, healthy retina were obtained using the fundus camera with a Spaide filter (a), confocal cSLO (b), and Optos ultrawidefield systems (c).
COURTESY OF BIOMED CENTRAL

To limit interference from the cornea and lens and reduce absorption from macular pigments, Richard Spaide, MD, introduced the “modified Topcon” filter set, with an excitation filter of 535-583 nm and barrier filter at 615-715 nm,²⁴ which captures red shifted wavelengths. In addition, while SW-FAF and fluorescein angiography share similar excitation spectra and cause cross-interference, the use of red-shifted wavelengths allows for concurrent use of the two techniques.²⁵

Studies have found that fundus cameras provide superior images in patients with exudative retinal disease, perifoveal RPE changes, and cataracts.^21,23,26 The limitations of fundus cameras include low contrast, interference by structures anterior to the retina, and the need to install filters after purchase.

Confocal Scanning Laser Ophthalmoscopy (cSLO)

Confocal scanning laser ophthalmoscopes use a series of mirrors to focus a low-power laser in a two-dimensional raster pattern onto the fundus (Figure 3B). Confocal optics then collect autofluorescent emissions from the retinal plane alone, eliminating lens and cornea artifacts.²⁷

Commercially available platforms such as the Heidelberg Spectralis (Franklin, MA) and the Nidek F-10 (Fremont, CA) use an excitation beam of 488 nm and a barrier filter of approximately 500-700 nm.²⁸ Real-time averaging produces high-contrast, high-resolution images, but in turn exposes the system to motion artifacts.²³

cSLOs also offer dual-wavelength autofluorescence, a technique used to calculate macular pigment density in the evaluation of retinal diseases.^3,29,30 Compared to fundus cameras, cSLOs produce superior images in 70% of cases.²³ However, cSLOs are more expensive, cannot be preceded by FA, and have no color options.

Optos Ultrawidefield FAF

The Optos Optomap (Marlborough, MA) ultrawidefield system combines confocal scanning laser technology and an ellipsoid mirror to provide 200º of view (Figure 3C). The system uses two excitation wavelengths of 633 nm (red) and 532 nm (green) and a barrier filter of >540 nm, which allow for concurrent use with FA.¹⁶

Ultrawidefield images are essential in documenting retinal conditions with peripheral findings.³¹ Limitations of the Optos system include distortion of the peripheral retina due to the ellipsoid mirror, lack of image averaging resulting in poor contrast, and lid/eyelash artifacts.³²

Additional ultrawidefield options include the Staurenghi lens, which provides 150º of view,³³ and the Heidelberg noncontact ultrawidefield system, with a 105º of view.³⁴

CONCLUSION

In the second part of this series, we will continue our discussion of FAF, including the characteristic findings in specific retinal conditions. RP

REFERENCES