Indocyanine Green Angiography in Age-Related Macular Degeneration

Indocyanine Green Angiography in Age-Related Macular Degeneration

Irene A. Barbazetto

Jason S. Slakter


Indocyanine green (ICG) angiography has evolved since the introduction of the technique and is used today to complement other imaging modalities such as spectraldomain optical coherence tomography (SD-OCT) and fluorescein angiography (FA). In neovascular age-related macular degeneration (AMD), ICG angiography is recommended for patients with serosanguineous macular detachments in the peripapillary area; serosanguineous macular detachments in the absence of drusen; large vascularized pigment epithelial detachments (PEDs), particularly with extensive lipid, blood, or minimal cystoid macular edema; hemorrhagic lesions in the peripheral retina; and lesions that are resistant or show suboptimal response to multiple anti-vascular endothelial growth factor (VEGF) indications and is based on the patients’ ethnicity (1).


ICG was introduced into medicine by Fox et al. (2) in 1957 as part of an indicator dilution technique for measuring cardiac output (3). Shortly thereafter, clinical applications were expanded, and ICG was utilized to quantify hepatic blood flow and assess liver function (4,5). It was the unique optical property of ICG that allowed for visualization of the dye through overlying melanin and xanthophyll (6) that made it interesting for ophthalmic imaging. Kogure and Choromokos were the first to image the choroid by injecting ICG dye into the carotid artery of monkeys. They were able to image choroidal veins but could not identify choroidal arteries or capillaries (7). In 1972, Flower and Hochheimer (8) were successful in imaging the choroidal circulation using ICG absorption angiography. One year later, technology evolved, and ICG fluorescence was imaged using ICG angiography, which allowed imaging of the choroidal arteries on a regular basis (9). The initial utility of ICG angiography was limited by the low fluorescence of the dye and recording techniques available, making imaging of the choriocapillaris difficult, if not impossible (10). The earliest studies using ICG angiography for AMD were done on film-based systems. Given the low fluence of the ICG dye, the detection of choroidal abnormalities was difficult. Patz et al. (10) were the first to describe the use of ICG angiography for choroidal neovascularization (CNV). They were able to identify choroidal lesions in only 2 of 25 patients. Other publications reported a similar low yield in regard to CNV detection (11,12,13). However, choroidal filling abnormalities were described in a significant subset of patients with advanced AMD (11). It took 20 years and
major advances in technology for this technique to become a standard diagnostic tool (14,15,16). Today, clinical images are recorded using digital angiography with a highly weighted camera in the near-infrared and high-speed imaging capacity (1,17) or scanning laser ophthalmoscope (SLO)-based technology (18,19,20). Together with FA, it has become a standard clinical tool for imaging choroidal pathologies such as AMD, central serous chorioretinopathy (CSC), uveitis, and choroidal tumors.


ICG (C43H47N2NaO6S2) is a tricarbocyanine dye that has a complex molecular structure with amphiphilic properties, meaning it has both hydrophilic and lipophilic characteristics. It appears to be rapidly and almost completely (98%) bound to plasma protein following its intravenous administration (5). Initially, it was thought that serum albumin was the main binding protein (5), but studies later suggested that 80% of ICG molecules actually bind to globulins such as alpha-1 lipoprotein (21,22,23). A recent study by Yoneya et al. (24) not only confirmed this finding but showed that ICG intensely binds to high-density lipoprotein (HDL) and moderately to low-density lipoproteins (LDLs). These lipoproteins have a large molecular size and may explain the limited vascular and tissue permeability of the dye (24).

The absorption and emission spectra of ICG are within the near-infrared range (21). ICG absorbs light between 790 and 805 nm and fluoresces between 770 and 880 nm, with a peak absorption or fluorescence at around 805 nm in plasma. The absorption spectrum is dependent on the solvent (saline solution vs. plasma and proteins) and the dye concentration, which tends to promote aggregate formation in higher concentrations and in the absence of protein binding (21,25). Clinically, the retinal pigment epithelium (RPE) and choroid absorb about 21% to 38% of the near-infrared light (800 nm) used in ICG angiography. By comparison, this is less than half of the amount of light absorbed during FA (500 nm) (6), thereby allowing for better visualization of choroidal pathologies. ICG fluence, however, reaches only 4% of the effective efficiency of fluorescein in the eye (26).

For clinical use, ICG is diluted with sterile water prior to injection with the exception of iodine-free ICG, which is dissolved into a 5% glucose solution. Diluted ICG is stable for 4 hours in plasma. Conversely, ICG in distilled water alone shows considerable decrease of optical density, which can be significantly accelerated by exposure to light (25).

Safety and Toxicity of ICG Angiography

ICG dye is well tolerated, and the rates of adverse reactions rank below those reported for FA. Severe adverse reactions were estimated at 0.05%, and the death rate following ICG angiography is approximately 1:333,333 (27).

Most commercially available ICG dyes contain about 5% sodium iodide. Therefore, the use of ICG is usually considered contraindicated in patients with iodine or shellfish allergies. Other relative contraindications include liver disease, uremia (11), and pregnancy (category C).


The introduction of digitized imaging in the 1990s allowed for bypassing some of the exposure problems experienced during the early experiments with ICG angiography due to the weak fluorescence emitted by the dye. Imaging systems used today are either modified fundus camera systems or laser scanning ophthalmoscopes. Regardless of the camera used, the basic technique remains the same: Prior to injection, infrared images are obtained. Twenty-five to fifty mg of ICG diluted in 2 mL aqueous solvent is injected intravenously followed by a 5-mL saline flush. Eight to ten seconds following the injection, rapid sequential photographs (1 per second) are acquired. When using videoangiography or SLO-based systems, short video segments (with a frame rate up to 30 images per second) can be recorded immediately after injection. It is important to start imaging prior to the dye appearance in the fundus in order to capture the earliest choroidal filling phase. Thereafter, images are obtained at 3- to 5-minute intervals for a total duration of 30 to 50 minutes. ICG angiography can be performed before, after, or simultaneously with FA.


ICG Angiography: Nonneovascular AMD

ICG angiography is not routinely used or recommended in nonneovascular AMD. Some studies, however, indicated that choroidal abnormalities in conjunction with drusen could be predictive of future progression to advanced AMD (28,29). Hanutsaha et al. (28) showed that in patients with unilateral CNV, the presence of focal hot spots or hyperintense plaques on ICG angiography in the fellow eye was associated with a significantly higher risk of progression to neovascularization (27% vs. 10% over an average of 22 months of follow-up). Pauleikhoff et al. (29) reported a study where a prolonged choroidal filling phase on ICG angiography increased the risk of geographic atrophy.


While angiography is not routinely indicated in the diagnosis of drusen, ICG studies have provided interesting details and thus enhanced our understanding of this pathology. Drusen fluorescence on ICG angiography varies with drusen size, location, and age. While hard drusen are hyper- or
isofluorescent, soft drusen usually remain hypofluorescent throughout the study (30,31). In younger patients who are more likely to have basal laminar drusen or malattia leventinese, drusen seem to be hyperfluorescent in the early and late phases of the ICG angiography (31). The exception are the large, aggregated drusen in the later stages of malattia leventinese, which are hypofluorescent in the early phases of ICG and present as hyperfluorescent spots surrounded by halos of hypofluorescence in the late phases (32).

Basal Laminar Drusen

Basal laminar drusen appear as minimally hyperfluorescent spots on early-phase ICG angiography. With time, they increase in fluorescence and become more confluent. However, a small subset of basal laminar drusen are described as being hypofluorescent in the late phase (33).

Reticular Drusen/Pseudodrusen/Subretinal Drusenoid Deposits

Reticular pseudodrusen, also referred to as subretinal drusenoid deposits, first described by Mimoun and associates in 1990, have been of interest because of their associated risk for developing advanced AMD (34,35,36). Querques et al. (37) described reticular pseudodrusen as hypofluorescent dots and dot pattern on ICG angiography, confirming observations reported by Arnold et al. (30) in an earlier publication. These patterns seemed to project on the choroidal stroma and follow but not overlay larger choroidal vessels.

ICG Angiography: Choroidal Neovascularization

Because of its specific properties, ICG angiography has been found to be especially useful in imaging CNV. ICG allows for imaging the near-infrared end of the spectrum, thereby enhancing the visualization of structures through the RPE, melanin, xanthophyll, blood, and serous fluid. As a proof of principle, histopathology of a patient with occult CNV confirmed that a late-staining, well-circumscribed, hyperfluorescent “plaque” seen on ICG imaging correlated precisely with a thin layer of fibrovascular tissue beneath the pigment epithelium (38).

Initially in the 1990s, ICG angiography was intended to expand the spectrum of lesions amenable to laser photocoagulation in patients with occult CNV. By using videoangiography, Yannuzzi et al. (17) were able to identify well-demarcated areas of CNV in 39% of patients previously classified as purely “occult” or type 1 neovascularization. Today, with the advent of anti-VEGF therapy, laser therapy has only a marginal role in managing AMD. The main purpose of ICG angiography in AMD has become to identify and monitor subgroups such as polypoidal choroidal vasculopathy (PCV) and retinal angiomatous proliferations (RAPs) or conditions that may mimic AMD like CSC.

Occult or Type 1 Choroidal Neovascularization

Guyer, Yannuzzi, and coworkers were the first to classify ill-defined CNV (occult or type 1 CNV) into subgroups based on their ICG angiographic appearance (27,39). The first group consists of patients presenting with a solitary, focal, and well-delineated area of hyperfluorescent CNV (hot spot). By definition, hot spots are less than 1 disk diameter in size. They represent hyperpermeable, active parts of the neovascular lesions and also include polyps in PCV and type 3 neovascularization in RAP and its final stage of chorioretinal anastomosis.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

May 22, 2016 | Posted by in OPHTHALMOLOGY | Comments Off on Indocyanine Green Angiography in Age-Related Macular Degeneration

Full access? Get Clinical Tree

Get Clinical Tree app for offline access