Fluorescein angiography is an indispensable procedure that can help you investigate the integrity of the retinal vasculature. Some common conditions for which fluorescein angiography is indicated include atrophic macular degeneration, central serous chorioretinopathy, cystoid macular edema, ischemic optic neuropathy and malignant choroidal melanoma. It also can help you diagnose diabetic retinopathy, retinal tumors, retinal vascular occlusions, rubeosis, subretinal neovascular membranes and unexplained vision loss.
This clinical review looks at the stages of testing and discusses how to interpret the results.
Stages of the Test
Fluorescein angiograms progress through five phases: pre-arterial (choroidal flush), arterial, arteriovenous, venous and late recirculation. The process begins with injection of sodium fluorescein dye into the antecubital vein. The dye reaches the posterior pole via the short posterior ciliary arteries, then spreads anteriorly to the ora serrata.
In most cases, choroidal filling begins 10-20 seconds after injection, and is first visible as a patchy and lobular pattern. Because the choroidal vascular system is fenestrated, fluorescein freely enters the extravascular spaces and appears as a generalized hyperfluorescence.
Choroidal flush appears as mottled fluorescence of the choriocapillaris, due to variable blockage by the retinal pigment epithelium. Choroidal fluorescence is normally absent in the macular area due to the presence of xanthophyll and lipofuscin pigments that absorb the cobalt blue exciting light.1
The arterial phase begins 1-2 seconds after choroidal filling, and is usually over within seconds of the first appearance of fluorescein dye. In a normal angiogram, branches fill simultaneously. Any delay in the filling of the arterial tree is abnormal and requires investigation.
In the arteriovenous phase, dye spreads to the pre-capillary arterioles, the capillaries and the postcapillary venules, resulting in a laminar flow pattern with a railroad track appearance in the filling venules.
Venous filling occurs in two phases: laminar and complete. In the laminar venous phase, you’ll see a flow pattern in which the blood closest to the vessel wall fluoresces. You’ll see a bright outline of fluorescence along the walls of the vessel and a dark line centrally in larger venules. This occurs due to the rapid flow of plasma along the vessel walls and the high density of erythrocytes in the central lumen.2 As more blood converges on the larger venules from smaller tributaries, complete venous filling occurs until the lumen completely fills with dye. The filling process usually takes 45-60 seconds. The maximum concentration of fluorescein within the choroid and retina occurs 20-25 seconds after injection. This point of the study, the peak phase, is the optimal time to see the foveal avascular zone. Shortly after, the dye begins to recirculate. You will see a gradual and progressive reduction in fluorescence as the kidneys remove dye from the bloodstream.
The late, or recirculation, phase takes about 7-20 minutes. By this time, arteries and veins are virtually devoid of fluorescein, and choroidal flush is barely perceptible. Photos taken at this time help identify late leakage of fluorescein, accumulation of intraretinal dye and staining of tissues with fluorescein.
Monitor the time it takes for the dye to travel from the arm to the eye, as well as filling times for the major retinal vascular branches. Next, identify areas of hypofluorescence or hyperfluorescence, and evaluate any change over the course of the study. As always, study the results to make sure they are consistent with your tentative diagnosis.
To interpret fluorescein angiography, measure the extent to which an area hypofluoresces or hyper-fluoresces relative to the normal background fluorescein pattern.
Defects that show hypofluorescence can fall into two categories: those that occur due to blockage, and those that happen because of impaired vascular filling. The best way to differentiate is to compare the angiogram with the ophthalmoscopic appearance of the retina. If the hypofluorescent area corresponds to a visible structure, then the hypofluorescence is likely due to blockage. If nothing is visible ophthalmoscopically, then nonperfusion of the area probably causes the hypofluorescence.2
When an overlying opaque substance shields the fluorescence beneath it, a blockage defect occurs. Conditions that cause this include hemorrhages, inflammatory cells, melanin, exudates, lipofuscin, fibrin, glial tissue, media opacities, congenital hypertrophy of the retinal pigment epithelium (CHRPE), or a condition that causes hyperpigmentation.4
Hemorrhages leading to hypo-fluorescence occur in one of these areas:
- Subretinal. These hemorrhages typically occur in degenerative or inflammatory conditions, and often result from a subretinal neovascular membrane.
- Intraretinal. Hemorrhages at this level often develop due to a weakening of the wall of a capillary or microaneurysm. Hemorrhages that originate from the venous end of capillaries are located in the compact middle layers of the retina, and have a dot-and-blot configuration.
Hemorrhages that originate from the more superficial precapillary arterioles follow the nerve fiber layer and have a characteristic flame-shape. These are commonly seen in hypertensive retinopathy and vein occlusions. They tend to block choroidal and retinal fluorescence in early views but may be difficult to see in late views because of the retinal edema sometimes observed with these conditions.
Intraretinal hemorrhages that arise from the vessels located in the inner nuclear layer block choroidal fluorescence without disrupting retinal fluorescence. Such intraretinal hemorrhages are common in diabetic retinopathy.
- Preretinal. These hemorrhages arise from the superficial capillary system or the retinal peripapillary system. Gravity causes the blood to settle, resulting in darker blood at the bottom of the D-shape or keel-shape. Preretinal hemorrhages often appear in proliferative diabetic retinopathy, and totally obscure fluorescence from both the choroid and retina.2
Hard exudates often result when macrophages attempt to remove lipid deposits from within the retina. They are located between the inner plexiform and inner nuclear layers, and usually indicate long-standing edema. The degree of hypofluorescence you see with hard exudates associated with diabetic retinopathy, Coat’s disease or exudative maculopathy depends on the density of the accumulated material. Diffuse or cystoid edema associated with diabetic retinopathy, vascular occlusive disease or post-op cataract complications may block fluorescence early in the angiogram, especially when the fluid is turbid.
Pigment hyperplasia from inflammatory disease occurs primarily at the RPE level. This hyperplasia results in blockage of choroidal fluorescence and is usually associated with adjacent areas of pigment loss (which will hyperfluoresce).
Pigment and scar tissue resulting from thermal injury in cryotherapy or photocoagulation also demonstrate hypofluorescence.5 However, scar tissue tends to hyperfluoresce in late views because of eventual fluorescein staining.2
Vascular filling defects fall into these three categories:
- Choroidal vascular filling defects. Either nonperfusion or a loss of choroidal tissue will manifest choroidal vascular filling defects. The clinical hallmark that will help you differentiate between the two is scleral visibility. When the choroid is absent, the sclera is visible on ophthalmoscopy. Choroidal tissue loss may result from congenital, hereditary or degenerative conditions, or trauma.6 The sclera is not visible in nonperfusion of choroidal tissue. Embolic occlusion may cause this. Usually, you’ll see a dark triangular area on the angiogram.
- Retinal filling defects secondary to vaso-occlusive disease. Arterial occlusions are associated with systemic disease, especially internal carotid and cardiac disease. They usually demonstrate nonperfusion of large areas of the retina in early angiography views. These areas are often edematous, and a plaque may sometimes be observed at an arteriolar bifurcation during ophthalmoscopy. Although some degree of patency will return to the occluded vessels, the infarcted area will often remain hypofluorescent in subsequent angiograms due to necrosis and atrophy of the retinal tissue.2
Branch or central retinal vein occlusions also have a strong association with systemic disease. A central vein occlusion commonly occurs at the lamina cribrosa, while a branch occlusion occurs most frequently at arteriovenous crossings and are usually the result of a thickened artery pressing on a thin-walled vein. A branch vein occlusion can occur as two distinct clinical entities: non-ischemic retinopathy and ischemic retinopathy. The clinical distinction is based on such factors as the degree of superficial hemorrhage, cotton-wool spots and associated arterial changes.7, 8 Blockage of the central retinal vein or a branched vein occlusion results in a classic retinal appearance of feathered hemorrhage, exudate and edema.
- Optic nerve filling defects. Capillary nonperfusion or an absence of tissue may cause filling defects in the optic nerve area. Optic atrophy appears dark on the angiogram because the capillaries are absent, while anterior ischemic optic neuropathy demonstrates hypofluorescence due to occluded sclerotic arterioles. In these cases, the optic disc is edematous and frequently surrounded by flame-shaped hemorrhages. The disc appears hypofluorescent in the early views, but can hyperfluoresce late in the study due to the accompanying retinal edema. Optic pits, frequently associated with serous elevation of the sensory retina, appear dark during early angiographic phases because of an absence of tissue.2
Three possible mechanisms cause hyperfluorescence: leakage of dye from an abnormal blood vessel; accumulation of dye in an abnormal space or staining of existing tissues; or greater visibility of dye due to an attenuation or loss of overlying structures.
A breakdown of the blood-retinal barrier causes leakage. It’s extravascular and most commonly appears late in the angiogram after blood vessels have emptied available dye. Typically, any early hyperfluorescence increases in size and intensity as the study progresses. When reading an angiogram, differentiate between leakage and staining. Staining occurs when fluorescein binds with tissue during its transit through the retina and choroid. Structures that frequently exhibit staining are the sclera, a healthy optic nerve, drusen and glial scars.
Lesions that produce retinal edema demonstrate late hyperfluorescence from the gradual accumulation of dye in extracellular spaces. Because of its architecture, retinal edema is commonly pronounced in the macular area. Macular edema may appear as a diffuse hyperfluorescence, or it may be cystic.
In cystoid macular edema, foveal capillaries leak fluid into the cystic spaces of the outer plexiform layer. It usually occurs secondary to other ocular or systemic conditions, and is mostly associated with cataract surgery with vitreous displacement. On angiography, this shows up as a classic petaloid (flower-like) pattern of hyperfluorescence.
When the edema is in the macular area, classify it as clinically significant if it shows hard exudates within 1/3 disc diameter (<500 microns) of the fovea with thickening of the adjacent retina; thickening of the retina within 1/3 disc diameter (<500 microns) of the fovea; or a zone of retinal thickening >1 disc area within 1 disc diameter of the fovea.9
Fluorescein pooling occurs in potential spaces such as those formed when the sensory retina detaches from the RPE, or when the RPE separates from Bruch’s membrane. Sensory retinal detachments usually result in a hyperfluorescent area with diffuse borders, while RPE detachments typically cause hyperfluorescent zones with distinct borders.
The hyperfluorescence observed with both subsensory and sub-RPE detachments persists very late in the angiogram. In central serous chorioretinopathy, a focal disruption in the RPE allows fluid to elevate the sensory retina and collect in the subsensory retinal space. This shows up as an early spot of focal fluorescence (inkblot pattern) that sometimes expands and takes on a mushroom-like (smokestack pattern) appearance as the dye permeates the subsensory retinal space.2
Abnormal vessels within the fundus include aneurysms, neovascularized tissue, arteriolar-venous shunts and telangiectatic vessels. Subretinal neovascularization associated with macular degeneration can be well delineated 1-2 seconds before the retinal arterioles fill and then leaks diffusely as the study progresses. Always be aware that associated hemorrhage often obscures subretinal neovascularization, making it difficult to identify.
If neovascularization exists, you’ll see it in early angiographic views as a lacy network of surface retinal vessels that fill rapidly and leak diffusely in late views. Retinal neovascularization is most commonly associated with proliferative diabetic retinopathy, but also exists in other retinal vascular conditions. In diabetic retinopathy, leakage can come from newly formed vessels that are more fragile than normal, or from pre-existing normal vessels that became inadequate. In either case, you end up with the development of microaneurysms that show up as hyperfluorescence on an angiogram.
Collateral or shunt vessels sometimes develop with diabetic retinopathy or after vascular occlusion. Fluorescein angiography is helpful when differentiating collateral (vein to vein) and shunt (artery to vein) vessels from neovascularization. Neovascularization typically leaks on angiography, while collateral or shunt vessels do not. Retinal telangeictasis is seen in congenital malformations, Coat’s disease, Leber’s miliary aneurysms and degenerative processes, and with epiretinal membrane formation. You can easily identify these vessels by their early hyperfluorescence.
Transmission or window defects occur when there is a loss or thinning of overlying tissue, allowing choroidal hyperfluorescence. Increased intensity of normal fluorescence is commonly caused by absence of the usual “filtering” pigment from the RPE. This can occur due to atrophy or inflammation and gives rise to a classic window defect.1 Transmission defects also occur with drusen, angiod streaks, pigment epithelial loss, geographic atrophy of macular degeneration, drusen, chorioretinal scars and full-thickness macular holes. 10
Fluorescein angiography is a valuable diagnostic tool that can help us better manage retinal disease. As we see more pathology in our roles as primary eye care providers, it will become increasingly important for us to be able to interpret, diagnose and make clinical decisions based on this test.
Dr. Gupta practices full-scope primary-care optometry at Stamford Ophthalmology. He can be reached at firstname.lastname@example.org.