(PDF 242 KB
With the introduction of fluorescein angiography decades ago, ophthalmologists observed that, even without the use of fluorescein, parts of the fundus showed areas of fluorescence in certain conditions. Although this faint, so-called autofluorescence was at first considered to be a distraction, it was later found to have potential as a diagnostic indicator and a tool for monitoring disease progression. New instrumentation and techniques have been developed to begin harnessing its potential.1
We now know that fundus autofluorescence (FAF) imaging can provide information about the health and function not just of the central retina but in the periphery as well.
The Fluorescent Fundus
What causes fundus autofluorescence? The retinal photoreceptors contain light-sensing molecules, a class of retinoids that are susceptible to damage and cross-linking, said Richard F. Spaide, MD, in private practice with Vitreous-Retina-Macula Consultants of New York. The photoreceptors shed their damaged outer segments, which the retinal pigment epithelium (RPE) ingests through phagocytosis. The molecules are stored in liposomes and form lipofuscin (LF).1
Accumulation of lipofuscin. “Over years, each RPE cell will eventually phagocytize 3,000,000,000 outer segments, and up to 25 percent of the cell volume will be occupied by lipofuscin,” said Dr. Spaide. In addition to age, disease states and potentially increased oxidative damage can contribute to the buildup of LF in RPE cells.
LF can be made to fluoresce by a fairly broad range of wavelengths, about 500 to 800 nm.1 “After excitation, the molecule will spontaneously release a photon of light of a somewhat longer wavelength, efficiently recorded with a variety of imaging methods,” said Dr. Spaide.
Hyper- and hypofluorescence. Areas of excess LF accumulation will appear hyperfluorescent. But when RPE cells die or are absent, LF disappears, leading to hypofluorescence.
Not just the RPE. Fluorescence comes from other layers of the retina as well. “With age and under certain pathologic conditions, this autofluorescence increases dramatically and adds to the signal produced by the lipofuscin in the RPE,” said Dr. Spaide.
Fundamentals of FAF
Two main methods are used to record autofluorescence: confocal scanning laser ophthalmoscopy (cSLO) and standard fundus camera photography, said Szilárd Kiss, MD, director of clinical research and assistant professor of ophthalmology at Weill Cornell Medical College in New York City.
cSLO systems. The first method developed for clinical FAF imaging, cSLO uses a laser to rapidly and precisely scan the retina, said SriniVas Sadda, MD, associate professor of ophthalmology at the University of Southern California and the Doheny Eye Center in Los Angeles. Among the cSLO systems equipped for FAF imaging are the Heidelberg Retina Angiograph, which uses blue light, and the Optos OCT
SLO, which uses green light.
You can acquire images at nearly a video rate and add them together, Dr. Sadda said, which improves the signal-to-noise ratio and provides higher-quality images. “More important, you’re able to image structures that may be relatively weak in terms of their autofluorescence.”
Because light is forced through the confocal pinhole, this method allows you to selectively isolate a signal from one plane only. “This method has the advantage of rejecting spurious fluorescence from other structures in the eye, such as the lens, which has many fluorophores and fluoresces in the same wavelengths as lipofuscin,” said Dr. Spaide.
However, because only one plane can be seen, he added, conditions that increase fluorescence from the neurosensory retina, including almost any condition causing a serous or tractional detachment, will not be detected well with a cSLO-based system.
Fundus camera. In contrast, “The fundus camera shows a summation of fluorescence from the fundus and consequently can image fluorescence from the retina and RPE at the same time,” said Dr. Spaide. “The signal strength is sufficiently powerful so that only one photograph needs to be taken, instead of averaging images over time.”
Based on modifications pioneered by Dr. Spaide, the fundus camera FAF-imaging method uses filters exciting in the green spectrum and recording emission in the yellow-orange spectrum, said Dr. Sadda.
Does color matter? Noting that some systems use blue and others green light for excitation, Dr. Kiss said, “These provide essentially corresponding information about the retina. However, the green spectrum may provide some additional detail in the fovea, as the blue light tends to be absorbed by the high concentration of xanthophyll pigments.” Dr. Sadda added, “One advantage of longer-wavelength [green] light is that there is less absorption by the crystalline lens of the eye, which is quite autofluorescent with blue light, especially in patients with cataracts.”
Near-infrared. Light outside of the visible spectrum can also be used. “Near-infrared FAF imaging uses very long wavelengths of light—about 790 nm—just beyond the ability of the human eye to detect,” said Dr. Sadda. These wavelengths excite molecules other than LF, most notably, melanin. The distribution of melanin in the eye, he said, provides a different type of information about the disease process than LF alone.
Wide-field imaging. FAF imaging can be used to view only the posterior pole or all the way out to the periphery of the retina. For the latter, “You need a wide-field imaging platform such as the Optos system, which has the capability of scanning to 200 degrees,” said Dr. Kiss. The Optos uses an ellipsoidal mirror in the image pathway, said Dr. Spaide, which deflects light across a wide field of the ocular fundus.
In a retrospective review, Dr. Kiss and colleagues found peripheral autofluorescent abnormalities in nearly 64 percent of eyes in patients with age-related macular degeneration (AMD), compared with about 36 percent of control eyes, suggesting potential implications for diagnosing and treating different subtypes of the disease.2 A retrospective analysis by Dr. Sadda found similar abnormal peripheral patterns in a majority of patients with a wide range of diseases.3
Because the Optos scans such a large area of the retina, however, it cannot capture images quite as rapidly as the Heidelberg, said Dr. Sadda. It is also less confocal, which may create some “noise” in the image due to autofluorescence from other structures.
Which Conditions to Test?
Unlike other imaging modalities, said Dr. Kiss, FAF provides functional information about retinal cells. According to Dr. Spaide, it is useful for almost any fundus disorder, including AMD, retinal detachment, inherited dystrophies, central serous chorioretinopathy, vitelliform lesions, and acute zonal occult outer retinopathy (AZOOR). Dr. Kiss noted that FAF is also helpful in screening for medication toxicity, including eye problems related to hydroxychloroquine (Plaquenil).
AMD. FAF imaging allows functional evaluation of geographic atrophy in dry AMD, said Dr. Kiss, making it possible to assess emerging therapies and monitor response to medication as well as progression of the disease.
“Areas of geographic atrophy are hypofluorescent because there are no photoreceptors and no lipofuscin within the RPE cells,” he said. However, hyperfluorescence shows up in junctional zones around geographic atrophy where the RPE is working overtime—a foreshadowing of imminent atrophy.
“Of the studies using autofluorescence as a secondary end point, the most prominent one is AREDS 2,” said Dr. Kiss. The results may have a strong effect on the application of FAF imaging in the future, he added, just as OCT now influences clinicians’ treatment of patients with wet AMD.
Retinal detachment. Hyperfluorescence in areas immediately adjacent to a retinal detachment can demarcate its extent and help explain visual problems in patients, said Dr. Kiss. “In a small series using ultra-wide-field FAF imaging, we also showed that these areas can remain abnormal, even after the retina is reattached.”
Dystrophies and degenerations. Retinal dystrophies and degenerations also show abnormal autofluorescence, said Dr. Kiss. As with dry AMD, retinal dystrophies such as retinitis pigmentosa demonstrate areas of both hyper- and hypofluorescence, a sign that the retina is burning out.
“Always test Best disease,” said Dr. Kiss, “which is characterized by macular or submacular accumulation of lipofuscin material and has a sharp, characteristic hyperfluorescent region in the macula.” Dr. Spaide added, “We are familiar with Best disease causing a yellow egg yolk–like lesion in kids and young adults, but with time that will go away. However, FAF imaging shows widespread abnormalities in the fundus and also can find areas of atrophy within the lesions.”4
Dr. Sadda finds FAF particularly useful in diagnosing Stargardt disease, in which pisciform lesions are readily apparent.
Central serous chorioretinopathy (CSC). According to Dr. Spaide, the extent of CSC is best seen with autofluorescence, not just in the area of the subretinal fluid but also in other parts of the macula or even the other eye.
As with diseases that cause a buildup of vitelliform material, said Dr. Spaide, CSC accumulates outer segments that have been shed but not yet phagocytized. “In central serous chorioretinopathy, it is common to see subretinal accumulation of material,” he said. “Some ophthalmologists will call that ‘subretinal protein’ without giving the material much thought. However, it is highly autofluorescent in the wavelengths used to excite retinoids. Proteins do not efficiently autofluoresce in these wavelengths, so the hypothesis that the material is protein does not fit the available facts.”
A Complementary Tool
The role of FAF imaging is still evolving, but it provides unique information to complement that obtained through other imaging modalities. With practice, it is possible to seamlessly integrate FAF imaging with other methods to gain a more complete understanding of a patient’s condition, said Dr. Spaide, just as someone learning tennis first individually practices serves, backhands, volleys, and forehand shots—and then uses them together fluidly in a tennis match.
FAF and fluorescein angiography. For example, both FAF imaging and fluorescein angiography provide useful and complementary information. “On the one hand, fluorescein angiography provides information about flow and the structure of blood vessels in ways autofluorescence can’t,” he said. “FAF imaging, on the other hand, captures fluorescence that is largely proportional to RPE lipofuscin accumulation and is a means to gauge certain aspects of physiology as well as topographic anatomic information.”
An overlooked method? Sometimes imaging modalities offer overlapping information, said Dr. Spaide, adding that, in such cases, it make sense to choose the one that is the least invasive, risky, or expensive. However, he noted, even seasoned clinicians may overlook FAF imaging—even though it’s an easy, safe, and cheap method that is particularly useful in establishing a difficult diagnosis.
1 Holz FG, Schmitz-Valckenberg S, Spaide RF, Bird AC. Atlas of Fundus Autofluorescence Imaging. Berlin: Springer-Verlag; 2007.
2 Witmer M et al. Acta Ophthalmol. 2012 May 11; doi: 10.1111/j.1755-3768.2012.02434.x.
3 Sadda SR et al. Widefield autofluorescence: a new tool to study macular and retinal disease. Presented at: AAO Retina Subspecialty Day; Oct. 22, 2011; Orlando, Fla.
4 Spaide RF et al. Ophthalmology. 2006;113(8):1392-1400.
Dr. Kiss is a consultant for and receives research funding from Optos. Dr. Sadda has been a consultant for Heidelberg Engineering, received royalties from Topcon Medical Systems, and received research support from Carl Zeiss Meditec, Optos, and Optovue. Dr. Spaide receives royalties from Optos and Topcon Medical Systems.