Fluorescence

Fluorescence is luminescence that is maintained only by continuous excitation. In other words, excitation at one wavelength occurs and is emitted immediately through a longer wavelength. Emission stops at once when the excitation stops. Fluorescence thus does not have an afterglow. A typical example of fluorescence is television. In the television tube, the excitation radiation is the electron beam from the cathode-ray tube. This beam excites the phosphors of the screen, which re-emit the beam as a glow that constitutes a television picture.

Sodium fluorescein is a hydrocarbon that responds to light energy between 465 and 490 nm and fluoresces at a wavelength of 520–530 nm. The excitation wavelength, the type that is absorbed and changed, is blue; the resultant fluorescence, or emitted wavelength, is green-yellow. If blue light between 465 and 490 nm is directed to unbound sodium fluorescein, it emits a light that appears green-yellow (520–530 nm).

This is a fundamental principle of fluorescein angiography. In the procedure, the patient, whose eyes have been dilated, is seated behind the fundus camera, on which a blue filter has been placed in front of the flash. Fluorescein is then injected intravenously. Eighty percent of the fluorescein becomes bound to protein and is not available for fluorescence, but 20% remains free in the bloodstream and is available for fluorescence. The blue flash of the fundus camera excites the unbound fluorescein within the blood vessels or the fluorescein that has leaked out of the blood vessels. The blue filter shields out (reflects or absorbs) all other light and allows through only the blue excitation light. The blue light then changes those structures in the eye containing fluorescein to green-yellow light at 520–530 nm. In addition, blue light is reflected off the fundus structures that do not contain fluorescein. The blue reflected light and the green-yellow fluorescent light are directed back toward the film of the fundus camera. Just in front of the film a filter is placed that allows the green-yellow fluorescent light through but keeps out the blue reflected light. Therefore the only light that penetrates the filter is true fluorescent light (Fig. 1.1).

Fig. 1.1 Absorption and emission curves of sodium fluorescein dye. The peak absorption (excitation) is at 465–490 nm (blue light). The peak emission occurs at 520–530 nm (yellow-green light).

Pseudofluorescence

Pseudofluorescence occurs when nonfluorescent light passes through the entire filter system. If green-yellow light penetrates the original blue filter, it will pass through the entire system. If blue light reflected from nonfluorescent fundus structures penetrates the green-yellow filter, pseudofluorescence occurs (Fig. 1.2). Pseudofluorescence (i.e., fake fluorescence) causes nonfluorescent structures to appear fluorescent. It can confuse the physician interpreting the fluorescein angiogram and lead him or her to think that certain fundus structures or materials are fluorescing when they are not. Pseudofluorescence also causes decreased contrast, as well as decreased resolution. Because fluorescein angiography uses black-and-white film, the nonfluorescent or pseudofluorescent light appears as a background illumination. The background illumination from pseudofluorescence is especially heightened if there are white areas of the fundus, such as highly reflective, hard exudates. Pseudofluorescence must be avoided. Therefore the excitation (blue) and barrier (green-yellow) filters should be carefully matched so that the overlap of light between them is minimal.

Fig. 1.2 Pseudofluorescence. The blue exciter filter overlaps into the yellow-green zone, and the yellow-green barrier filter overlaps into the blue zone. The combination results in pseudofluorescence.

Film-based versus digital fluorescein angiography – historical perspectives

Fluorescein angiography finds its origins in the late 1960s with the publication of an original article describing its use as well as subsequent atlases and textbooks for a medical retinal specialty in its infancy.^1,6 The landmark text Atlas of Macular Diseases by Dr J. Donald Gass set a new standard for the use of stereoscopic fluorescein angiography in fundus diagnosis.⁷

As digital photography has evolved with improved resolution, the convenience of digital-based fluorescein angiography has gained wider acceptance. Though film-based images offer the highest amounts of resolution and 35-mm negatives are often easier to view for stereo, images are relatively difficult to manipulate, and training and effort are required to process and duplicate film. Transmitting or sharing film-based images is also time-consuming compared with digital images (Box 1.1).⁸

Camera and auxiliary equipment

Cameras differ in the degree of fundus area included in the photographs. Fundus cameras may range from 35° to 200° wide-field camera systems such as Optos.^9,10 In clinical retinal practice, cameras ranging from 35° to 50° are routinely used (Fig. 1.3). Regardless of range, a camera with the ability to yield high resolutions of the posterior pole is essential for most macular problems, especially when laser treatment is to be done, as with background diabetic retinopathy, branch-vein occlusion, or choroidal neovascularization.

Fig. 1.3 Digital fundus camera for color fundus and fluorescein angiography.

Wide-angle angiography has the benefit of capturing a single image of the retina in high resolution well beyond the equator. The potential for clinical efficiency and sensitivity in detecting neovascularization in the far periphery as well as acquiring an excellent clinical picture of the degree of capillary retinal nonperfusion is an exciting development in fluorescein angiography (Fig. 1.4).

Fig. 1.4 Optos wide-field images. (A) Nonperfusion detected in the left eye with wide-field fluorescein imaging. (Courtesy of Umar Mian, MD. Image taken by Carolina Costa.) (B, C) Wide-field angiography of the right and left eyes of a patient with diabetic retinopathy. Note the multiple areas of leakage corresponding to areas of retinal neovascularization associated with capillary nonperfusion. It is often difficult in certain wide-field images to determine the presence of small neovascular complexes versus leakage from capillary nonperfusion, unless areas are magnified further.

(Courtesy of Szilárd Kiss, MD.)

Light sources (viewing bulb and flash strobe)

In film-based cameras, bulbs burn out, but are easily replaced. A supply of each should always be kept on hand (Fig. 1.5).

Fig. 1.5 Fundus film camera with side wall removed to view inside parts. Yellow arrow, viewing bulb; black arrow, flash bulb; red arrow, exciter filter wheel. The light from the flash goes through the system in this photograph from right to left. The light is reflected off a mirror and travels upward to another mirror; it is then reflected to the left, into the patient’s eye. From there it is reflected directly to the right of the fluorescein camera back (white arrow).

Fluorescein solution

Sodium fluorescein, an orange-red crystalline hydrocarbon (C₂₀H₁₂O₅Na), has a low molecular weight (376.27 Da) and readily diffuses through most of the body fluids and through the choriocapillaris, but it does not diffuse through the retinal vascular endothelium or the pigment epithelium.

Solutions containing 500 mg fluorescein are available in vials of 10 mL of 5% fluorescein or 5 mL of 10% fluorescein. Also available are 3 mL of 25% fluorescein solution (750 mg). The greater the volume, the longer the injection time will be; the smaller the volume, the more likely a significant percentage of fluorescein will remain in the venous dead space between the arm and the heart (see Injecting the fluorescein, below). For this reason we prefer 5 mL of 10% solution (500 mg fluorescein).

Fluorescein is eliminated by the liver and kidneys within 24 hours, although traces may be found in the body for up to a week after injection. Retention may increase if renal function is impaired. The skin has a yellowish tinge for a few hours after injection, and the urine has a characteristic yellow-orange color for most of the first day after injection.

Various side-effects and complications can occur with fluorescein injection (Box 1.2).^11–15

A serious complication of the injection is extravasation of the fluorescein under the skin. This can be extremely painful and may result in a number of uncomfortable symptoms. Necrosis and sloughing of the skin may occur, although this is extremely rare. Superficial phlebitis also has been noted. A subcutaneous granuloma has occurred in a few patients after fluorescein extravasation. In each instance, however, the granuloma has been small, cosmetically invisible, and painless. Toxic neuritis caused by infiltration of extravasated fluorescein along a nerve in the antecubital area can result in considerable pain for up to a few hours. The application of an ice pack at the site of extravasation may help relieve pain. For extremely painful reactions an injection of a local anesthetic at the site of extravasation is effective but rarely necessary.

If extravasation occurs, the physician must decide whether to continue angiography. Extravasation may occur immediately, and thus the serum concentration of the fluorescein will be insufficient for angiography. In this case it usually is best to place the needle in another vein and reinject a full dose of fluorescein, starting the process again from the beginning. Occasionally, only a small amount of fluorescein is extravasated at the end of the injection. In this case photography can continue without stopping or reinjecting.

A common cause of extravasation is the use of a large, long needle directly attached to a syringe. It is difficult to hold the syringe in the dark. For this and other reasons we have discussed earlier, a scalp-vein needle attached to a syringe by a flexible tube is the best choice for this procedure. Also, the patient’s own blood can be drawn back into the tubing of the scalp-vein needle, with the blood going all the way up to but not into the syringe. When it is time to inject, the person giving the injection can look at the tip of the needle to ensure that extravasation has not occurred. If it has, the patient’s own blood is extravasated, and little chance of complication exists if the injection is stopped at this point so that no fluorescein is injected.

It is always important to watch for extravasation at the beginning of the injection so that, should it occur, the process can be halted; thus only a minimal amount of fluorescein will have been injected and extravasated. The amount of extravasated fluorescein can be minimized by slow injection and constant observation of the needle with a hand-held light or if injection is done before turning off the room lights.

Nausea is the most frequent side-effect of fluorescein injection, occurring in about 5% of patients. It is most likely to occur in patients under 50 years of age or when fluorescein is injected rapidly. When nausea occurs, it usually begins approximately 30 seconds after injection, lasts for 2–3 minutes, and disappears slowly.

Vomiting occurs infrequently, affecting only 0.3–0.4% of patients.^11,13 When it does occur, it usually begins 40–50 seconds after injection. By this time most of the initial-transit photographs of the angiogram will have been taken. A receptacle and tissues should be available in case vomiting does occur. When patients experience nausea or vomiting, they must be reassured that the unpleasant and uncomfortable feeling will subside rapidly. Photographs can be taken after the vomiting episode has passed. A slower, more gradual injection may help to prevent vomiting.

Patients who have previously experienced nausea or vomiting from fluorescein injection may be given an oral dose of 25–50 mg of promethazine hydrochloride (Phenergan) by mouth approximately 1 hour before injection. Promethazine has proved to be helpful in preventing or lessening the severity of nausea or vomiting. We have recently found that we can also reduce the incidence of nausea by warming the vial of fluorescein to body temperature and drawing it into the syringe through a needle with a Millipore filter. Restriction of food and water for 4 hours before the fluorescein injection may reduce the incidence of vomiting; an empty stomach may prevent vomiting but will not affect nausea. If the patient still has a tendency to vomit despite taking all these measures, a lesser amount of fluorescein can be given and injected more slowly if the photographic results will not be compromised.

Vasovagal attacks occur much less frequently during fluorescein angiography than does nausea and are probably caused more by patient anxiety than by the actual injection of fluorescein. We have seen vasovagal attacks even when the patient sees the needle or immediately after the skin has been penetrated by the needle but before the injection has begun. Occasionally a vasovagal reaction causes a patient to faint, but consciousness is regained within a few minutes. If early symptoms of a vasovagal episode are noted, smelling salts usually reverse the reaction. The photographer must be alert for signs of fainting because the patient could be injured if he or she falls.

Shock and syncope (more severe vasovagal reaction) consist of bradycardia, hypotension, reduced cardiovascular perfusion, sweating, and the sense of feeling cold. If the photographer and person injecting see that the patient is getting “shocky” or lightheaded, the patient should be allowed to bend over or lie down with the feet elevated. The patient’s blood pressure and pulse should be carefully monitored. It is important to differentiate this from anaphylaxis, in which hypotension, tachycardia, bronchospasm, hives, and itching occur.

Hives and itching are the most frequent allergic reactions, occurring 2–15 minutes after fluorescein injection. Although hives usually disappear within a few hours, an antihistamine, such as diphenhydramine hydrochloride (Benadryl), may be administered intravenously for an immediate response. Bronchospasm and even anaphylaxis are other reactions that have been reported, but these are extremely rare. Epinephrine, systemic steroids, aminophylline, and pressor agents should be available to treat bronchospasm or any other allergic or anaphylactic reactions. Other equipment that should be readily available in the event of a severe vasovagal or anaphylactic reaction includes oxygen, a sphygmomanometer, a stethoscope, and a device to provide an airway. The skilled photographer observes each patient carefully and is alert to any scratching, wheezing, or difficulty in breathing that the patient may have after injection.

There are a few published and unpublished reports of death following intravenous fluorescein injection. The mechanism may be a severe allergic reaction or a hypotensive episode induced by a vasovagal reaction in a patient with pre-existing cardiac or cerebral vascular disease. The cause of death in each case may have been coincidental. Acute pulmonary edema following fluorescein injection has also been reported.

There are no known contraindications to fluorescein injections in patients with a history of heart disease, cardiac arrhythmia, or cardiac pacemakers. Although there have been no reports of fetal complications from fluorescein injection during pregnancy, it is current practice to avoid angiography in women who are pregnant, especially those in the first trimester.

Aligning camera and photographing

To align the fundus camera properly, the photographer must first assess the “field of the eye.” The camera is equipped with a joystick with which the photographer can adjust the camera laterally and for depth. The camera is also equipped with a knob for vertical adjustment. The photographer finds the red fundus reflex, which is an even, round, sharply defined, pink or red light reflex. If the camera is too close to the eye, a bright, crescent-shaped light reflex appears at the edge of the viewing screen or a bright spot appears at its center. If the camera is too far away, a hazy, poorly contrasted photograph results.

The photographer moves the camera from side to side to ascertain the width of the pupil and the focusing peculiarities of the particular cornea and lens. The photographer studies the eye through the camera lens, moving the camera back and forth and up and down, looking for fundus details (e.g., retinal blood vessels). The photographer then determines the single best position from which to photograph (Figs 1.6 and 1.7).

Fig. 1.6 The patient’s arm rests on an adjustable armrest that is elevated so that the patient’s arm is at or above the level of her heart. The armrest also facilitates easy placement of the intravenous needle and injection of fluorescein.

Fig. 1.7 The patient’s head is kept steady in the chinrest and headrest of the fundus camera. The photographer aligns the camera and focuses on the patient’s right fundus. Each is in a comfortable position, facilitating the stability necessary to achieve a good fluorescein angiogram.

Occasionally, a patient has a peculiar corneal reflex or central lens opacity, and it may be impossible to follow the usual procedure of aligning the camera through the central axis of the eye. Moving the camera slightly off axis may help improve focus and resolution.

Any abnormalities, such as an unusual light reflex or a poorly resolved image the photographer sees through the camera system, will appear on the photograph. If the ophthalmoscopic view seen through the camera is not optimal, the photograph will not be optimal (Fig. 1.8). If the view is optimal, well aligned, in focus, and without reflexes, the photograph can be optimal. A helpful concept for the photographer is “what you see is what you get (or worse – never better).”

Fig. 1.8 Fundus photograph and reflexes. (A) Photograph of right fundus without reflexes. The camera was properly aligned and focused. (B) Note the bright red, yellow and blue arc on the right side. The flash is reflecting off the iris. This can be remedied by repositioning the camera slightly to the right or left. (C) In this case the camera was placed at the proper distance from the fundus but was placed too far to one side (down and to the right), which allowed the bright white arc reflex to the lower right. (D) Note white reflex, especially above, in, and below the papillomacular bundle. In this case the camera was in proper alignment but was placed too far away from the patient’s eye.

Focusing

Achieving perfect focus is a major factor in the photographic process. Both the eyepiece crosshairs and the fundus details must be in sharp focus to obtain a well-resolved photograph. The proper position of the eyepiece is determined by the refractive error of the photographer and the degree to which he or she accommodates while focusing the camera.

The photographer first turns the eyepiece counterclockwise (toward the plus, or hyperopic, range) to relax his or her own accommodation; this causes the crosshairs to blur. The photographer then turns the eyepiece slowly clockwise to bring the crosshairs into sharp focus. The eyepiece is focused properly when the crosshairs appear sharp and clear (Fig. 1.9, online). They must remain perfectly clear while the photographer focuses on the fundus with the camera’s focusing detail. With experience, the photographer becomes expert in adjusting the eyepiece and in keeping the crosshairs in focus throughout the entire photographic sequence.

Fig. 1.9, online The photographer focuses the eyepiece of the camera by initially turning the eyepiece counterclockwise, then clockwise, and stopping when it is in exact focus. The photographer must be sure that the eyepiece crosshairs remain in perfect focus throughout the photographic procedure.

The best position for the eyepiece is the point at which the crosshairs are in focus while the photographer’s accommodation is relaxed. Photographers learn to relax accommodation by keeping both eyes open. The photographer focuses the eyepiece with one eye and, with the other eye, keeps a distant object, such as the eye chart, in sharp focus. This skill may be difficult for technicians without ophthalmic training, but it is seldom impossible to learn.

Keeping the crosshairs in sharp focus, the photographer then turns the focusing dial on the camera to focus the fundus detail. Some photographers focus the crosshairs just once at the beginning of each day and control their accommodation throughout the day. This is not a good idea because the photographer’s accommodation may change during a photographic session; the photographer should be aware of this possibility and regularly check and readjust the eyepiece for focus. With the camera properly aligned and focused, the photographer is ready to start the preliminary photographs and angiograms.

Using stereophotography

Stereophotography separates, photographically, the tissues of the eye for the observer. Stereo fluorescein angiography facilitates interpretation by separating in depth the retinal and choroidal circulation.^16,17 Stereo angiography is considered absolutely essential in certain situations.¹⁸ The photographic protocol for the Macular Photocoagulation Study required stereo fluorescein angiography. Without well-resolved stereo images, interpretation of angiograms with, for instance, choroidal neovascularization associated with age-related macular degeneration, can be extremely difficult, if not impossible (Fig. 1.10, online). On the other hand, stereophotography, although extremely helpful in cases that are difficult to interpret, is not always absolutely necessary because other fundus features and characteristics usually indicate the level at which abnormal fluorescence is located.

Fig. 1.10, online Viewing stereo fundus photographs. (A) Negatives are placed on a viewing back-lit display. Two negative images are then viewed with an adjustable stereo lens. Reading the contact print of the angiogram. The stereo viewer can be easily made up using a trial frame. In viewing negative images, “hyperfluorescence” corresponds to dark objects while corresponding “hypofluorescent” objects are lighter. (B) Reading a stereo pair of digital angiograms. The special viewer allows the observer to focus both images. The software displayed is Ophthalmic Imaging Systems.

Adequate stereophotographs can be achieved with a pupillary dilation of 4 mm, although dilation of 6 mm or more is best. The first photograph of any stereo pair is taken with the camera positioned as far to the photographer’s right (the patient’s left) of the pupil’s center as possible (of course, without inducing reflexes). The second photograph of the stereo pair is taken with the camera held as far to the photographer’s left (the patient’s right) of the pupil’s center as possible. This order is extremely important because the photographs are taken and positioned on the film so that the angiogram is read from right to left. Thus the first photograph in the stereo sequence appears on the right on the contact sheet to correspond with the interpreter’s right eye; the second is printed on the left for his or her left eye. It follows, then, that the first view of a stereo pair should be taken from the photographer’s right, followed by a view from the left.

Photographing the periphery

Photographing the peripheral retina with a standard 50° fundus camera demands precision and skills acquired only after many hours of practice. Problems with patient position and camera alignment and focus are compounded by marginal corneal astigmatism, unsteadiness of patient fixation, light reflexes, and awkward camera placement. All steps necessary for taking posterior photographs, such as alignment and focusing, must be employed to achieve good peripheral fundus photography. The Zeiss camera comes with an astigmatic dial to help neutralize the induced astigmatism. A tilt mechanism, now standard on most cameras, helps position the camera for extreme superior and inferior peripheral photography (Fig. 1.11).

Fig. 1.11 Photographing the periphery. (A) The tilt mechanism of the camera allows the back of the camera to be lifted up (tilted to aim downward) for photography of the inferior periphery. The same tilt mechanism can be used to bring the camera far down (tilted to aim upward) to take pictures of the superior periphery (B). In photographing the inferior periphery, the photographer must sometimes stand. This photograph was a mock situation. In a real situation, the photographer or an assistant would have to lift the patient’s upper lid to view the inferior periphery properly.

During photography of the periphery, the patient tends to turn or move his or her head. Unsatisfactory photographs caused by the movement of the head away from the camera or to the side can be avoided if the photographer is alert to these possibilities and takes the necessary steps to prevent them. On the whole, achieving good peripheral photographs depends on photographic skill, of course, but also on patience on the part of both photographer and patient.

Informing the patient

An important step toward a successful angiogram is to inform the patient about the procedure. An informed patient is generally less anxious and more cooperative than one who is unsure of the situation. Some institutions routinely provide a consent form to be signed by patients who are to have angiograms. However, this practice cannot replace the duty of the physician to inform the patient about the procedure and its potential complications and to answer all questions.

The patient should be told that the eyes will be dilated, sodium fluorescein will be injected in a vein in the arm or back of the hand, and photographs will be taken. The patient should be assured that the flash is a harmless, bright light (not an X-ray) and that fluorescein dye is safe. The patient should be told that injection of the dye can cause complications but that such occurrences are rare. If the patient requests further details about complications, the physician is obligated to supply the information.

Positioning the patient

Before the patient is seated at the camera, the photographer makes sure that the front lens is free of any dirt or dust. The lens should always be covered by a lens cap when the camera is not in use. The front of the lens should be kept clean using chloroform and a tightly rolled rod of lens tissue. To clean the lens, begin at the center and rotate out to the periphery.

The patient is positioned at the camera with the chin in the chinrest and the forehead against the head bar. Because the most common cause of poor fluorescein photographs is involuntary movement of the patient’s head, the photographer should prepare and make adjustment for this before the fluorescein is injected. The photographer should aim and focus the camera on the specific area of primary interest, at the same time noting the patient’s responses. If the photographer finds that the camera must continually be moved closer to the patient while aligning it or taking preliminary photographs, or if reflexes suddenly appear in the view even though the camera is steady, then the patient’s head has moved away from the chinrest. If so, the photographer can make some adjustment before injecting the fluorescein dye. Sometimes having an assistant hold the patient’s head in the chinrest is helpful (Fig. 1.12, online). The photographer either may lower the entire camera and chinrest or raise the patient’s chair. This causes the patient to lean forward in the chinrest and against the forehead bar, making it more difficult for the patient to pull back.

Fig. 1.12, online An assistant holds the patient’s head as a reminder to the patient to keep the chin in the chinrest and forehead against the bar.

Before photography begins, and between shots, the photographer may ask the patient to blink several times. This usually makes the patient more comfortable and also moistens the cornea and keeps it clear. When the pictures are actually being shot, the patient should be instructed to blink as infrequently as possible.

The photographer should talk to the patient frequently during the procedure, informing the patient of the progress of the testing and assuring him or her that all is going well. Explanation and reinforcement help produce better photographs.

Injecting the fluorescein

The color stereoscopic fundus photographs are taken first, before the fluorescein is injected. For injection, we recommend a syringe with a 23-gauge scalp-vein needle (Fig. 1.13, online). The scalp-vein needle has several advantages: it is small enough to enter most visible veins, and an intravenous opening is then available in the event of an emergency. Once in the vein, it requires no further attention, and although it can be taped in place, this usually is unnecessary. Whenever an antecubital vein is not visible or accessible, the vein in the back of the hand or radial (thumb) side of the wrist can usually be used for injection. Injecting the fluorescein into a hand or wrist vein increases the circulation time by a few seconds, but this seldom makes any difference.

Fig. 1.13, online Ten percent fluorescein solution, 5 mL syringe, and 23-gauge scalp-vein needle.

Injection of the fluorescein is coordinated with the photographic process and is done after the first photographs (color fundus and control photographs; see next section) have been taken. With the needle in place, angiography can begin. By a predetermined, preferably silent, signal (such as a nod of the head), the photographer indicates to the physician to begin injecting fluorescein. The photographer starts the timer on the camera simultaneously with the start of injection and takes one photograph. This frame will show zero time on the photograph. In this way, the time from the beginning of injection is recorded on each subsequent angiographic photograph. When the injection is finished, the photographer may take another picture, which shows how long the injection took.

A rapid injection of 2 or 3 seconds delivers a high concentration of fluorescein to the bloodstream for a short time and probably yields somewhat better photographs than does a slower injection. However, the more rapid the injection, the greater the incidence of nausea from a highly concentrated bolus of fluorescein. For this reason a slower injection (4–6 seconds) is preferable; the photographs will still be of good quality. Because some fluorescein dye remains in the tubing, the scalp-vein needle should have short, rather than long, tubing to ensure that more of the dye is injected.

Developing a photographic plan

To photograph and print the fluorescein angiogram, we suggest the following comprehensive plan, designed to yield maximal angiographic information from each fundus and to facilitate a thorough and complete interpretation (Fig. 1.14). In contrast to film-based angiograms which required multiple duplicate attempts to assure at least one image would be optimal, fewer photographs are necessary in digital angiography. Although most angiograms will be complete by following this procedure, there will be exceptions. This plan must be modified if abnormalities occur in areas other than the macula and disc.

Fig. 1.14 Photographic plan for fluorescein angiography of macular disease. Film-based images printed upside down because the fundus camera inverts the image of the fundus, and, to read the angiogram upright, the film is printed with the frame numbers upside down. In digital photography, no inversion is required.

The photographic strategy essentially begins when the clinician has identified a condition or finding that requires angiographic study. The pathology dictates whether the photographic approach should image a magnified highly detailed finding versus a wider field of view for a more diffuse disease. Narrower field limits with higher magnification yield optimal images for focal pathology in conditions such as maculopathies, optic nerve disorders, and small focal lesions. Wider field of view often sacrifices magnification, but is effective in documenting conditions involving the periphery, such as diabetic retinopathy and vascular occlusive disease. Peripheral retinal scans for areas of neovascularization, as well as consideration for images that can be later montaged for wide-field reproductions, must also be incorporated into certain photographic plans. Elevated lesions such as tumors require great care in capturing high-quality stereo images.

It is both essential and extremely cost-effective for the physician to indicate specifically what areas to photograph. The photographer should be directed as to where to start the angiogram and the issues important for each specific angiogram. It is most efficient to use a photographic instruction slip that indicates the specific number of color photographs to take of each area, where to start the angiogram, what the diagnosis is, and any other information about the patient or fundus that is pertinent to the photographic process (Fig. 1.15, online). Although digital color and angiograms avoid the issue of wasted film and developing costs, unnecessary computer storage of images and patient inconvenience can be avoided with good technique and a repeatable, accurate algorithm for angiography.

Fig. 1.15, online Photographic request form. In the top left portion of the form, the physician indicates the number of color photographs required for each area in the fundus. The physician also indicates the diagnosis because experienced photographers will know which type of photographs to take for each particular diagnosis. The physician also indicates in which location of the fundus the initial-transit phase of the angiogram should take place or, in other words, where the photographer should start the angiogram. In the lower right portion of the photographic request form, the physician can indicate to the photographer specifics about the patient that will facilitate the photographic process and save the photographer time. The physician can indicate whether an eye can fixate on a light, whether the media are clear (so that if the photographer cannot get a clear view, he or she can immediately understand that it is caused by a problem of the eye), and what the patient’s refraction is so that the photographer can know which special lenses to use in the photographic process.

Historically, because the roll of 35-mm negative film used for fluorescein angiography has 36 frames, it was convenient to think of the photograph session in terms of six rows of six frames each. Thus frame 1 appears in the upper right-hand corner and frame 36 in the lower left-hand corner. The angiogram developed from 36-mm negatives thus reads from right to left and from top to bottom. With the advent of digital imaging, theoretically, an unlimited number of frames can be acquired. However, to maximize efficiency of resources, digital storage of 20 frames per digital proof sheet is typically more than adequate for most clinical scenarios.

The first frame of the angiographic series is the color photograph of each eye. Then, a preinjection “control” photograph checks the dual-filter system for autofluorescence and pseudofluorescence.

At this point the fluorescein injection is begun. The needle is inserted in a vein in the patient’s arm (Fig. 1.16, online). The photographer waits for confirmation of successful venous access and awaits verbal confirmation that infusion is about to begin. Once the injecting clinician starts the infusion of fluorescein, the photographer begins the initial “injection” image. When the injecting clinician has completed infusion, he or she announces “injection complete” and the photographer takes the “end-of-injection” image. Because it is important to observe the site of the needle tip for extravasation of fluorescein, the lights are turned off only at the end of the injection. An alternative method is to turn the lights off after the needle has been inserted in the vein. The person injecting can hold a hand light to observe the fluorescein flow into the vein to be sure extravasation is not occurring. With the lights off, the photographer can become dark-adapted, which allows him or her to be better able to see the flow of fluorescein into the fundus as it occurs.

Fig. 1.16, online After the needle is placed in the vein, the lights can be turned off so that the photographer can become dark-adapted and see fluorescein flow in the eye. With the use of a hand light, the person injecting can carefully observe the injection site so as to be sure extravasation is not occurring. In this way the fluorescein solution can be injected while the room lights are out.

So as not to miss the appearance of fluorescein as it enters the fundus, the photographer should begin taking the initial-transit fluorescein photographs 8 seconds after the beginning of the injection of the dye if the patient is young and 12 seconds after injection for older patients. This is done so that these early photographs will not miss the appearance of fluorescein as it enters the fundus. Then, at intervals of 1.5–2 seconds, approximately six photographs should be taken in succession.

If the photographer does not see fluorescein entering and filling the retinal vessels while the six initial-transit photographs are taken, he or she must continue to photograph the fundus until filling takes place and also should check to see why no fluorescein is present.

After the first six initial-transit photographs and approximately 20–30 seconds after injection, with sufficient fluorescein concentration in the eye, the photographer should take a photograph of the fellow eye, a stereoscopic pair of photographs of the primary area of interest, followed by a stereoscopic pair of other pertinent areas. For example, in the suggested photographic plan, after stereophotographs of the right macula are taken, stereophotographs of the right disc are taken. The photographer should then photograph in stereo the macula and disc of the fellow eye.

Late-stage angiographs, preferably in stereo, are taken of the pertinent areas of each eye. It is important to photograph both discs and macula and any other areas of abnormal fluorescence and to note any areas that could not be photographed. This ensures that the interpreter will have adequate information for a complete interpretation of the angiogram.

The entire photographic process lasts 5–10 minutes. Angiophotographs taken more than 10 minutes after injection are usually not necessary. In some cases of central serous chorioretinopathy, or other rare situations, angiophotographs taken longer than 10 minutes after injection are helpful. This photographic algorithm is modified for specific conditions. For instance, in an angiogram of a diabetic patient, peripheral scans may be included at the request of the clinician surveying for neovascularization. In a patient with possible choroidal neovascularization due to age-related macular degeneration, additional angiophotographs of the suspicious lesion may be useful.

At the end of the session the patient is asked regarding any sensations related to an allergic reaction and reminded that the urine will be discolored for about a day.

In the event of a technical difficulty, such as camera breakdown, repeat fluorescein injection or photography can be carried out with satisfactory results after a waiting period of 30–60 minutes.

The plan we have suggested allows the fluorescein angiogram to yield all the information necessary to make a proper and thorough interpretation.

Box 1.3 provides a checklist of important steps in the fluorescein angiography procedure.

Fundus anatomy and histology

Fluorescein angiography has greatly increased our knowledge of retinal and choroidal circulatory physiology and fundus pathology. This clinical and research tool facilitates the in vivo study of histopathologic characteristics of fundus disease. Before the advent of fluorescein angiography, conditions such as pigment epithelial detachment, cystoid retinal edema, and subretinal neovascularization could be evaluated and understood only histologically. Now they are widely appreciated and recognized clinically. Because fluorescein angiography graphically demonstrates fundus pathophysiology, and because we rely on histologic points of reference to interpret a fluorescein angiogram, a thorough knowledge of the anatomy of the fundus and its microscopic layers is necessary to interpret fluorescein angiograms correctly. To interpret a fluorescein angiogram, it is essential to understand the microscopic layers of the fundus (i.e., the histology).

A logical place to begin this study is at the vitreous. In its normal state, and in a normal angiogram, the vitreous is clear and nonfluorescent. However, when it contains opacities that block the view of retinal and choroidal fluorescence, hypofluorescence occurs. The vitreous is also an important point of reference when intraocular inflammation or retinal neovascularization is present. In these cases fluorescein leaks into the vitreous, causing fluffy fluorescence as fluorescein molecules disperse into fluid vitreous and vitreous gel.

For the purpose of fluorescein angiographic interpretation, it is convenient to divide the sensory retina into two layers: the inner vascular half and the outer half, which is avascular. The inner vascular half extends from the internal limiting membrane to the inner nuclear layer. This portion of the retina contains the retinal blood vessels, which are located in two separate planes: the larger retinal arteries and veins are located in the nerve fiber layer; the retinal capillaries are located in the inner nuclear layer. In a well-focused stereoscopic fluorescein angiogram, these two vascular layers can be seen as distinct planes in the retina. An extremely important fluorescein angiographic concept is that normal retinal blood vessels are impermeable to fluorescein leakage; that is, fluorescein flows through the normal retinal vessels without leakage into the retina.

The outer avascular half of the sensory retina comprises the outer plexiform layer, the outer nuclear layer, and the rods and cones. The outer plexiform layer is the primary interstitial space in the retina. When the retina becomes edematous, it is in this layer that fluid accumulates, causing cystoid spaces. Deep retinal hemorrhages and exudates (lipid deposits) may also be deposited in the outer plexiform layer.

The rods and cones are very loosely attached to the pigment epithelium, especially in the macular region, whereas the pigment epithelium is very firmly attached to Bruch’s membrane. In fluorescein angiographic interpretation the pigment epithelium is an extremely important tissue because it prevents fluorescein leakage from the choroid and blocks, to a greater or lesser extent, visualization of choroidal fluorescence.

Bruch’s membrane separates the pigment epithelium from the choriocapillaris, which is permeable to fluorescein. Fluorescein passes freely from the choriocapillaris and diffuses through Bruch’s membrane up to, but not into, the pigment epithelium. Beneath the choriocapillaris are the larger choroidal vessels, which are impermeable to fluorescein. Melanocytes are dispersed throughout the choroid but are most heavily concentrated in the lamina fusca, the thin layer between the choroid and sclera. The sclera lies beneath the choroidal vessels.

The ophthalmic artery gives rise usually to two main posterior ciliary arteries: the lateral and medial. However, three posterior ciliary arteries may be present, in which case the medial artery is the one usually duplicated less frequently. In rare instances there may be a superior posterior ciliary artery.

The posterior ciliary arteries supply the lateral and medial halves of the disc and choroid. During angiography a vertical zone of slightly delayed filling may be seen passing through the papillomacular region, including the disc. Occasionally, there is an oblique orientation to this supply or even a superoinferior distribution. This border between the main posterior ciliary arteries has been termed the watershed zone, where patchy choroidal filling often can be seen on fluorescein angiograms.

Each main posterior ciliary artery divides into numerous short arteries and one long artery. On the temporal side the short posterior ciliary arteries supply small, variously sized, wedge-shaped choroidal segments, whose apices are centered near the macula. The lateral long posterior ciliary artery passes obliquely through the sclera. It supplies a wedge of choroid that begins temporal to the macular region and participates in the formation of the greater circle of the iris.

The choriocapillaris is made up of discrete units called lobules, thought to be approximately one-fourth to one-half of a disc diameter in size. The center of each lobule is fed by a precapillary arteriole (terminal choroidal arteriole), which comes from a short posterior ciliary artery. Each lobule functions independently in the normal state. It has been assumed that angiographic zones of delayed or patchy choroidal filling gradually fill in a transverse fashion, with one lobule spilling over into another. Careful inspection, however, indicates that these filling defects generally remain the same size, indicating a delayed filling from a posterior origin (its own arteriolar feeder). In the abnormal state, as when a choroidal vascular occlusion occurs, there is a freely connecting “spilling over” of blood flow from well-perfused choroid to the occluded area.

Around the margin of each lobule is a ring of postcapillary venules that drain each lobule. These postcapillary venules drain into the vortex veins, which drain the entire choroid. There are usually four vortex veins, and each functions as a well-defined quadrantic segmental drainage system for the entire uvea. In the case of a posterior ciliary artery obstruction, this occluded portion of the choroid can fill by a retrograde mechanism from an adjacent posterior ciliary artery by way of the choroidal venous system. This mechanism may provide adequate nourishment to prevent extensive ischemic changes until the occluded artery reopens.

Knowledge of each of these layers of the fundus is important in understanding fundus histopathology. The following six areas, however, are more important than others in the interpretation of abnormal fundus fluorescence:

Throughout this chapter a modified schematic drawing relates various fluorescein angiographic abnormalities to fundus histopathologic changes (Fig. 1.17). The size and proportion of these various layers have been modified to include various pathologic manifestations and to illustrate the effects of these abnormalities on the angiogram. Because of its importance and various pathologic changes, the pigment epithelium is drawn to a larger scale in relation to other fundus structures. Only the inner portion of the sclera is represented because the outer portion of the sclera is usually of little importance to angiographic interpretation. The retinal and choroidal vessels are drawn larger and more numerous than they appear in a normal histopathologic section to emphasize the contribution of circulatory pathophysiologic interpretation.

Fig. 1.17 Modified schematic drawing of a microscopic section of retina, pigment epithelium, and choroid.

Two specialized areas of the fundus warrant more detailed discussion: the macula (Fig. 1.18) and the optic nerve head. The fovea is the center of the macula and contains only four layers of the retina: (1) the internal limiting membrane; (2) the outer plexiform layer; (3) the outer nuclear layer; and (4) the rods and cones. No intermediate layers exist between the internal limiting membrane and the outer plexiform layer in the fovea, which in the macula is oblique. This is an important factor in understanding the stellate appearance of cystoid edema in the macula as opposed to the honeycomb appearance of cystoid edema outside the macula. Beyond the macular region the outer plexiform layer is perpendicular rather than oblique.

Fig. 1.18 Modified schematic drawing of a microscopic section of the macula.

The pigment epithelial cells in the macula are more columnar and have a greater concentration of melanin and lipofuscin granules than in the remainder of the fundus.

Xanthophyll is present in the fovea, located probably in the outer plexiform layer. These differences in pigmentation are the chief factors responsible for producing the characteristic dark zone in the macular region on normal angiograms. The absence of retinal vessels in the fovea (i.e., the perifoveal capillary-free zone), in most cases approximately 400–500 mm in diameter in the center of the fovea, is another cause of the dark appearance of the macula.

The optic nerve head, or disc, is the other highly specialized tissue of the posterior pole. The disc is fed by two circulatory systems: the retinal vascular system and the posterior ciliary vascular system. Widespread anastomotic channels exist between the posterior ciliary vasculature and the optic nerve and retinal vasculature and become exaggerated in certain pathologic conditions. The disc is made up of many layers of nerve fibers and glial supporting columns that contain the large retinal vessels.

The central retinal artery arises from the ophthalmic artery in close proximity to the main posterior ciliary arteries. In about 45% of the population, the central retinal artery and the medial posterior ciliary artery arise from a common trunk. In 12% of persons the central retinal artery originates from the ciliary artery. Therefore it is impossible to have a choroidal infarction, anterior ischemic optic neuropathy, and a central retinal artery occlusion all due to a single site of obstruction.

The central retinal artery provides a major source of blood supply to the axial portion of the anterior orbital portion of the optic nerve. In the intraneural or axial course, short centrifugal branches arise but usually end a short distance behind the lamina cribrosa. There are then no further branches from the central retinal artery until it reaches the retina. If a cilioretinal artery is present, it supplies the corresponding segment of the disc.

The peripapillary nerve fiber layer is supplied by small, recurrent branches from the retinal arterioles at the peripapillary region. Emanating from these arterioles at the disc are the radial papillary capillaries. These capillaries are rather straight and long, have few anastomoses, and lie in the superficial portion of the peripapillary nerve fiber layer. The capillaries to the disc are continuous with these retinal peripapillary capillaries.

The short posterior ciliary arteries, or the recurrent branches from the peripapillary choroid, supply the retrolaminar portion of the optic nerve. The laminar cribrosa portion of the nerve is supplied by centripetal branches of the short posterior ciliary arteries. In this region a partial, or, rarely, a complete Zinn’s vascular circle is occasionally found. The prelaminar portion is supplied by centripetal branches from the peripapillary choroid.

Because most of the disc is fed by the ciliary system, fluorescein appears simultaneously at the optic nerve head and the choroid and before it is apparent in the retinal arteries.

The main venous drainage of the disc is into the central retinal vein. The prelaminar portion empties into both the central retinal vein and the peripapillary choroid, thus providing potential collateral drainage in the case of obstruction of the central retinal vein behind the lamina cribrosa. Such large dilated collaterals are frequently seen following central retinal vein occlusion and are called retinociliary veins. Some mistakenly call them opticociliary shunts, a misnomer because they are not true shunts (defined as a congenital artery that empties into a vein and that skips the capillary bed, sometimes part of the Wyburn–Mason syndrome), and they are not optico because they emanate from the retina. They are, most accurately, retinovenous to ciliovenous collaterals.

In summary, fluorescein angiography provides an in vivo understanding of the histopathologic and pathophysiologic changes of various fundus abnormalities. Therefore an anatomic and, more specifically, a histologic understanding of important fundus landmarks is essential to fluorescein angiographic interpretation.

Normal fluorescein angiogram

The normal fluorescein angiogram is distinguished by certain specific characteristics. Knowledge of these characteristics provides an essential frame of reference for interpreting abnormal fluorescein angiograms.

In the normal fluorescein angiogram (Fig. 1.19), the first true fluorescence begins to show in the choroid approximately 10–12 seconds after injection in young patients (e.g., adolescents) and 12–15 seconds after injection in older patients.

Fig. 1.19 Normal fundus photos and fluorescein angiogram of left disc and macula taken with a 50° camera. (A) Montage photograph of multiple fields shows normal macula, fovea, and retinal vessels. (B) Early arterial phase of the fluorescein angiogram. Note the ground-glass fluorescence of the choriocapillaris. There is very little fluorescence in the retinal veins; just the margins of the veins are fluorescent. This is the earliest portion of the laminar filling phase of the vein. Note some hyperfluorescence of choriocapillaris. These dark patches of the choroid are areas that have not fully filled, referred to as patchy choroidal filling. (C) The retinal arteries and capillaries have filled and the retinal veins have filled more substantially. Note the laminar flow in the retinal veins; this is indicated by the white line of fluorescence along the walls of the retinal veins. (D) Late venous phase. Laminar filling is no longer detectable and uniform filling is seen in both arterial and venous circulation. (E) Mid to later arteriovenous phase of fluorescein angiogram. Note that the ground-glass fluorescence of the choriocapillaris is complete. The retinal arteries and veins are completely filled. (F) Arteriovenous phase of fluorescein angiogram showing the disc. Again, there is diffuse fluorescence of the choriocapillaris. The arteries and veins have completely filled, and optic nerve fluorescence is normal.

Fluorescence can appear even earlier than 8 seconds in very young patients. The choroid occasionally begins to fluoresce 1 or 2 seconds before the initial filling of the central retinal artery. Early choroidal fluorescence is faint, patchy, and irregularly scattered throughout the posterior fundus. It is interspersed with scattered islands of delayed fluorescein filling. This early phase is referred to as the choroidal flush. When adjacent areas of choroidal filling and nonfilling are quite distinct, the pattern is designated as patchy choroidal filling.

Within the next l0 seconds (approximately 20–25 seconds after injection), the angiogram becomes very bright for about 5 seconds because of the extreme choroidal fluorescence. Choroidal fluorescence, however, is not visible in the macula because of the taller, more pigmented epithelium present in the fovea (retina). Therefore the macula remains dark throughout the angiogram.

If present, a cilioretinal artery usually begins to fluoresce as the choroid fluoresces, rather than as the retina fluoresces. Within 1–3 seconds after choroidal fluorescence is visible, or approximately 10–15 seconds after injection, the central retinal artery begins to fluoresce. The less dense the concentration of pigment in the pigment epithelium, the greater the time between the visibility of the choroidal fluorescence and the filling of the retinal vessels. The lighter pigment presents less interference to choroidal fluorescence, allowing it to be evident earlier in its filling phase. With a more densely pigmented pigment epithelium, the blockage barrier effect is greater. Therefore choroidal fluorescence appears somewhat later because a greater concentration of fluorescein is required to overcome the increased density of the pigment epithelial barrier.

Because no barrier exists in front of the retinal vessels, the patient’s pigmentation has no effect on the visibility of the retinal vessels, although the degree of pigmentation does affect the contrast of the angiophotographs. The darker the pigment epithelium is, the less visible the choroidal fluorescence will be and the greater the contrast of the retinal vascular fluorescence (i.e., the better they stand out). The lighter the pigment epithelium is, the more visible the choroidal fluorescence will be and the less the contrast of the fluorescence from the retinal vessels.

After the central retinal artery begins to fill, the fluorescein flows into the retinal arteries, then into the precapillary arterioles, the capillaries, the postcapillary venules, and finally the retinal veins. Because the fluorescein from the venules enters the veins along their walls, the flow of fluorescein in the veins is laminar. Because vascular flow is faster in the center of a lumen (tube) than on the sides, the fluorescein seems to stick to the sides, creating the laminar pattern of retinal venous flow. The dark (nonfluorescent) central lamina is nonfluorescent blood that comes from the periphery, which takes longer to fluoresce because of its more distant location.

In the next 5–10 seconds, fluorescence of the two parallel laminae along the walls of the retinal veins becomes thicker. At the junction of two veins, the inner lamina of each vein may merge. This creates three laminae: one in the center and one on each side of the vein. As fluorescein filling increases in the veins, the laminae eventually enlarge and meet, resulting in complete fluorescence of the retinal veins.

Fluorescence of the disc emanates from the posterior ciliary vascular system, both from the edge of the disc and from the tissue between the center and the circumference of the disc. Filling also comes from the capillaries of the central retinal artery on the surface of the disc. Because healthy disc tissue contains many capillaries, the disc becomes fairly hyperfluorescent on the angiogram.

The perifoveal capillary net cannot always be seen on the fluorescein angiogram. It can be seen best in young patients with clear ocular media about 20–25 seconds after a rapid fluorescein injection. This is called the “peak” phase of the fluorescein angiogram. The photographer should be aware of this phase and be sure not to miss it by shooting as rapidly as possible as the fluorescein concentration increases and by continuing to shoot rapidly until the concentration of fluorescein begins to decrease.

Approximately 30 seconds after injection, the first high-concentration flush of fluorescein begins to empty from the choroidal and retinal circulations. Recirculation phases follow, during which fluorescein in a lower concentration continues to pass through the circulation of the fundus.

Generally, 3–5 minutes after injection, the choroidal and retinal vasculatures slowly begin to empty of fluorescein and become gray. Vessels of most normal patients almost completely empty of fluorescein in approximately 10 minutes. The large choroidal vessels and the retinal vessels do not leak fluorescein. However, because of large gaps in its endothelium, the choriocapillaris does leak fluorescein. The extravasated fluorescein diffuses through the choroidal tissue, Bruch’s membrane, and sclera. Leakage of fluorescein with retention in tissues is designated as staining. In the later phase of the angiogram, staining of Bruch’s membrane, the choroid, and especially the sclera may be visible if the pigment epithelium is lightly pigmented. The disc and adjacent visible sclera remain hyperfluorescent because of staining. When the retinal pigment epithelium is especially lightly pigmented, the large choroidal vessels can be seen in silhouette against the fluorescent (fluorescein-stained) sclera. The lamina cribrosa within the disc also remains hyperfluorescent because of staining. This depends on the cup-to-disc ratio and the presence of any visible sclera, such as occurs within a conus adjacent to the disc. The edge of the disc stains from the adjacent choriocapillaris, which normally leaks.

To summarize, the angiogram is initially dark; choroidal and retinal filling is seen 10–15 seconds after fluorescein injection. The retinal and choroidal vasculatures fill maximally about 20–30 seconds after injection. Late angiophotographs show fluorescence of the choroid and sclera (if the pigment epithelium is light) and fluorescence of the optic cup and the edge of the disc, but otherwise the fundus is dark (nonfluorescent in the late phase).

Hypofluorescence

Hypofluorescence is a reduction or absence of normal fluorescence, whereas hyperfluorescence is abnormally excessive fluorescence. A systematic series of decisions follows this initial differentiation to arrive at a proper diagnosis. These decisions relate to: (l) the anatomic location of various abnormalities; (2) the quality and quantity of the abnormal fluorescence; and (3) other unique characteristics, as indicated in Fig. 1.20.

Hypofluorescence is any abnormally dark area on the positive print of an angiogram. There are two possible causes of hypofluorescence: blocked fluorescence or a vascular filling defect.

Blocked fluorescence is sometimes referred to as masked, obscured, or negative fluorescence or transmission decrease. Each of these terms indicates a reduction or absence of normal retinal or choroidal fluorescence because of a tissue or fluid barrier located anterior to the respective retinal or choroidal circulation. For example, blood in the vitreous or a layer of blood in front of the retina obscures the view of the retinal and choroidal circulations and therefore blocks fundus fluorescence from these tissues. Hemorrhage that lies under the retina or retinal pigment epithelium, but in front of the choroidal circulation, does not obstruct visibility of the retinal circulation but does block the view of the choroidal circulation. Therefore the approximate histologic location of blocking material can be determined by the presence or absence of visibility of one or both fundus circulations.

Fluorescein is present but cannot be seen in blocked fluorescence. With vascular filling defects, however, fluorescein cannot be seen because it is not present.

The key to differentiating blocked fluorescence from a vascular filling defect is to correlate the hypofluorescence on the angiogram with the ophthalmoscopic view. If there is material visible ophthalmoscopically that corresponds in size, shape, and location to the hypofluorescence on the angiogram, then blocked fluorescence is present. If there is no corresponding material on the color photograph, then it must be assumed that fluorescein has not perfused the vessels and that the hypofluorescence is caused by a vascular filling defect.

Hypofluorescence resulting from a vascular filling defect occurs when either of the two fundus circulations is not perfusing normally. This is caused by an absence of the vascular tissue or by a complete or partial obstruction of the particular vessels. In these situations an absence or delay of fluorescence of the involved vessels will occur. This type of hypofluorescence has a pattern that follows the geographic distribution of the vessels involved. Although the ophthalmoscopic picture will demonstrate the material blocking fluorescence, it may show nothing if the hypofluorescence is the result of a vascular filling defect.

To summarize, after an area of hypofluorescence is recognized, one must refer to the ophthalmoscopic photograph to determine the cause. If material is visible ophthalmoscopically and corresponds to the area of hypofluorescence, this is blocked fluorescence. If no corresponding blocking material exists, the hypofluorescence is therefore a vascular filling defect.

Blocked retinal fluorescence

Blocked retinal vascular hypofluorescence is caused by anything that reduces media clarity. An opacification in front of the retinal vessels involving the cornea, anterior chamber, iris, lens, vitreous, or the most anterior portion of the retina or disc produces hypofluorescence.

The further the opacification is in front of the fundus, the less it will block fluorescence and the more it will affect the overall quality of the photographs. The closer the material is to the fundus, the more it will block, causing hypofluorescent images on the angiogram. Any material that blocks retinal vascular fluorescence will, of course, block choroidal fluorescence as well.

Any anterior-segment material, such as a corneal opacity, anterior-chamber haziness, or lens opacity, obscuring the view of the ocular fundus will result in an angiogram of reduced brilliance, contrast, and resolution. This affects the quality of the angiogram and is, in a sense, a type of blocked fluorescence.

Many conditions of the vitreous produce a hazy medium that prevents visualizing fundus detail. The most common vitreous opacity to cause blockage is hemorrhage. Whether diffusely dispersed in the vitreous gel or more densely accumulated, vitreous hemorrhage reduces or completely blocks fundus fluorescence. In addition to hemorrhage, media haze may be caused by a variety of opacifications, including asteroid hyalosis, vitreous condensation resulting from vitreous degenerative disease, inflammatory debris, vitreous membranes, or opacification secondary to amyloidosis. When anterior-segment and vitreous opacities are present, the angiogram may be of higher resolution and quality than the color photograph because the light scattered from the nonfluorescing opacities is not transmitted through the barrier filter and therefore has no effect on the angiographic photograph.

Any translucent or opacified material in the retina or in the nerve fiber layer blocks fluorescence from both planes of retinal vessels, as well as from the choroidal vessels. The large retinal vessels and precapillary arterioles are located in the nerve fiber layer in the anterior plane of the retina. The capillaries and postcapillary venules are located deeper in the retina, in the inner nuclear layer. If a blocking material lies in front of the nerve fiber layer, it blocks both planes of retinal vessels (Fig. 1.21). However, if the material lies beneath the nerve fiber layer but within or in front of the inner nuclear layer (where the smaller retinal vessels are located), it blocks only the retinal capillaries (and choroidal vessels), leaving the view of the large retinal vessels unobstructed. If a blocking material lies deeper than the retinal vascular structures, deep to the inner nuclear layer, it does not block the vessels but will block the choroidal vascular fluorescence. In other words, deep intraretinal blocking material, such as hemorrhage or exudate, does not obstruct retinal vascular fluorescence, since the retinal vessels are located in the inner half of the retina (Fig. 1.22).

Fig. 1.21 Preretinal hemorrhage causing hypofluorescent blockage of all retinal and choroidal fluorescence. (A) Schematic drawing of subhyaloid (right), subinternal limiting membrane (central), and nerve fiber layer (left) hemorrhages. Each hemorrhage lies in front of the retinal, and therefore choroidal, vasculature, causing hypofluorescence-blocked fluorescence. (B) Color photograph of the right disc showing substantial preretinal hemorrhage. (C) Fluorescein angiogram of the right disc showing hypofluorescence caused by blockage as a result of the preretinal hemorrhage. Comment: All fluorescence of the fundus is blocked because the hemorrhage lies in front of the retinal vasculature.

Fig. 1.22 Intraretinal hemorrhages causing hypofluorescent blockage. (A) Schematic of retina showing hemorrhages located in most of the layers of the retina from the internal limiting membrane to the outer nuclear layer. (B) Color photograph of left macula shows dot-and-blot, as well as flame-shaped hemorrhages just above the fovea. This is a case of branch-vein occlusion. (C) Fluorescein angiogram of left macula shows that the hemorrhage causes irregular hypofluorescent blockage. The flame-shaped hemorrhage located in the nerve fiber layer blocks all the retinal vasculature. The dot-and-blot hemorrhages do not block the large retinal vessels and therefore can be localized deeper in the retina. The hemorrhages that do not block retinal capillary fluorescence can be located deeper to the capillary layer, which is in the inner nuclear layer. Comment: Once hypofluorescent blockage is determined, an anatomic localization of the blocking material can be made by determining which normally fluorescent structures can be seen and which are being blocked.

Therefore one can determine the location of a retinal abnormality, such as hemorrhage, by the vessels that are blocked by it and by the fluorescence of the vessels that are not blocked.

The most common cause of blocked retinal vascular fluorescence is hemorrhage. Subinternal limiting membrane hemorrhage blocks fluorescence of all underlying retinal vessels and choroidal vasculature. Nerve fiber layer hemorrhage, which usually is flame-shaped, blocks the smaller retinal vessels lying deeper in the retina but only partially blocks the larger retinal vessels in the nerve fiber layer. Blockage from hemorrhage is usually complete, as opposed to the partial blockage caused by the myelinated nerve fibers.

Various retinal vascular (arteriolar) occlusive diseases may cause white ischemic thickening (nerve fiber edema), which results in some opacification of the retina and blockage of the remaining retinal vascular and choroidal fluorescence. Conditions such as arterial occlusion in hypertension or Purtscher’s retinopathy cause enough intracellular “cloudy” swelling and opacification to block fluorescence. It should be noted that, because there is occlusion in this type of hypofluorescence, the hypofluorescence is caused partly by the vascular filling defect. However, the opacified ischemic retina effectively blocks fluorescence from underlying retinal and choroidal vasculature.

In summary, the concept of blocked retinal vascular hypofluorescence is fairly easy to understand and to identify on the angiogram. When the retinal vessels do not fluoresce, the ophthalmoscopic view should be studied to determine if blocking material is located in front of the retinal vessels.

If blocking material is present, the next step is to determine its anatomic location.

Blocked choroidal fluorescence

Hypofluorescence caused by blocked choroidal vasculature occurs when fluid, exudate, hemorrhage, pigment, scar, inflammatory material, or the like accumulates in front of the choroidal vasculature and deep to the retinal vasculature (Fig. 1.23).

Fig. 1.23 Subretinal hemorrhage causing hypofluorescence, specifically, blockage of choroidal fluorescence. (A) Schematic of retina with subretinal hemorrhage (blood located between photoreceptors and pigment epithelium). (B) Color photograph of right macula of an eye with angioid streaks showing large scattered areas of subretinal hemorrhage. (C) Fluorescein angiogram of right macula shows marked hypofluorescence caused by blocked choroidal fluorescence (the retinal vessels are visible) that is due to the subretinal hemorrhage. Comment: The subretinal hemorrhage completely obscures fluorescence from the choroid. The retinal vessels are clearly seen overlying the subretinal hemorrhage.

Deep retinal material

Materials deposited in the deep retina that cause blockage of choroidal fluorescence are fluid, hard exudate, hemorrhage, and pigment.

Fluid that accumulates in the deep retina has a predilection for the tissue of least resistance, the outer plexiform layer. Deposition of edema fluid, originating from leaking retinal vessels or migrating from subretinal space into the retina, most frequently occurs in the outer plexiform layer. After reaching a certain volume, the fluid tends to form spaces, or pockets, between compressed nerve and Müller’s fibers, which are pushed aside in the process. This pattern of fluid accumulation in the outer plexiform layer is called cystoid retinal edema. Noncystoid retinal edema occurs when the volume of extracellular fluid is insufficient to produce pockets, or spaces, in the outer plexiform layer or other layers of the retina. A significant amount of retinal edema, whether cystoid or noncystoid, especially if turbid or containing lipid-laden macrophages, partially blocks choroidal fluorescence in the early phase of the fluorescein angiogram. Later in the angiogram, retinal edema fluoresces. Intraretinal hard exudates and lipid-laden macrophages, usually located in the outer plexiform layer, partially block choroidal fluorescence. When retinal vessels bleed, the blood can be deposited anywhere in the retina. When located deep to the retinal vessels beneath the inner nuclear layer, retinal vascular fluorescence is visible, whereas choroidal fluorescence is blocked.

Subretinal material

Any opaque or translucent substance located beneath the retina but in front of the choroid blocks fluorescence of the choroidal vasculature but does not block retinal vascular fluorescence (Fig. 1.23). Blood located under the retina causes complete blockage of choroidal fluorescence, with the retinal fluorescence showing through normally. Subretinal hemorrhage appears red, and subpigment epithelial hemorrhage is dark. Subretinal hemorrhage is generally scalloped with somewhat irregular margins, whereas subpigment epithelial hemorrhage is often quite round and well demarcated (Fig. 1.23).

Accumulated pigment (melanin and lipofuscin) from diseased retinal pigment epithelium causes blocked choroidal fluorescence (Fig. 1.24). Any hyperpigmentation of the pigment epithelium causes blocked choroidal fluorescence. Xanthophyll, the pigment present in the outer layers of the fovea, blocks choroidal fluorescence by selectively absorbing the blue exciting light, which results in less fluorescence. Finally, a choroidal nevus may block much of the choroidal fluorescence (Fig. 1.25) and especially blocks the later hyperfluorescent staining of the sclera. The choriocapillaris may be seen normally over the nevus.

Fig. 1.24 Hypertrophy of the retinal pigment epithelium. (A) Schematic showing hypertrophic pigment epithelial cells. (B) Color photograph of the macula shows a well-demarcated hyperpigmented lesion. (C) Fluorescein angiogram of the same area shows marked hypofluorescence of the choroid resulting from blocked fluorescence. Comment: This patient had marked hypertrophy of the retinal pigment epithelium, which allowed normal retinal fluorescence; it completely blocked choroidal fluorescence.

Fig. 1.25 Choroidal nevus hypofluorescent blockage. (A) Schematic drawing of retina showing choroidal nevus. Note that the choriocapillaris is intact. (B) Color photograph of nevus. (C) Arteriovenous phase of fluorescein angiogram shows hypofluorescence corresponding to the area of the nevus. (D) Later arteriovenous phase of fluorescein angiogram shows that the nevus is still hypofluorescent, although the choriocapillaris ground-glass fluorescence can be seen surrounding.

To summarize, various materials located in the deep retinal layers, or beneath the retina, block choroidal fluorescence and are evident ophthalmoscopically. These materials result from a variety of disease processes.

Vascular filling defect

The second cause of abnormal hypofluorescence is vascular filling defect. With blocked fluorescence, the fluorescein is present in the circulations of the fundus but is not visible because a tissue or fluid barrier conceals it. With vascular filling defect, fluorescein cannot be seen because it is not present. Since fluorescein reaches the retina and choroid by way of vessels, lack of the fluorescein dye in either vascular system indicates an obstructive problem or a lack of vessels (i.e., a vascular filling defect).

As previously indicated, when a hypofluorescent area is seen on an angiogram, the best way to differentiate blocked fluorescence from a vascular filling defect is to compare the angiogram with the ophthalmoscopic picture. When blood, pigment, or exudate can be seen ophthalmoscopically corresponding to the area of hypofluorescence, the material is causing blocked fluorescence. When no material is visible ophthalmoscopically (on the color photograph), one must assume that fluorescein has not perfused the vessels and that the abnormal hypofluorescence is caused by a vascular filling defect. In some instances both forms of hypofluorescent mechanisms play a role simultaneously, as with retinal arteriolar occlusion, when the retina is not only not perfused (vascular filling defect) but is ischemic and therefore white and opaque, causing blocked fluorescence.

Vascular filling defects result from vascular obstruction, atrophy, or absence (congenital or otherwise) of vessels. Any of these conditions can be total or partial. When the obstruction is complete (occlusion) or the vascular tissue is atrophied completely, the hypofluorescence is complete and lasts throughout the angiogram. When the obstruction is only partial or the vascular tissue is not entirely atrophied, the vascular fluorescein filling is delayed or reduced relative to corresponding areas that fill normally. Whatever the cause of a partial vascular filling defect, hypofluorescence will be seen in the early phases of the angiogram but may not persist throughout the entire angiogram. Some vascular filling, although delayed or reduced, will eventually occur.

Once it is determined that a vascular filling defect is the cause of an area of hypofluorescence, the next step is to determine which of the retinal, disc, or choroidal vessels are involved. A vascular filling defect of the disc is easy to discern angiographically. Determining whether a vascular filling defect is retinal or choroidal can be more difficult. Since retinal vessels are normally present, however, the absence of retinal vessels is usually readily apparent. If, on the other hand, a vascular filling defect is found but the retinal vessels are full and visible, the hypofluorescence must be choroidal in origin. Stereoscopic angiophotographs allow one to distinguish between the planes of the retina and choroid and enable exact determination of the location of the hypofluorescence.

Retinal vascular filling defect

If a retinal vascular filling defect is present, the clinician then considers whether the defect results from obstruction of a retinal artery or vein, capillary bed, or any combination of these. Distinguishing the cause of the obstruction is not difficult because the fluorescein angiographic process is dynamic and timed. When nonfilling of a specific retinal vessel occurs, it is easy to differentiate an arterial occlusion from a venous occlusion because the retinal arteries fill first, then the retinal capillary bed, followed by the retinal veins. In addition, retinal vascular filling defects can be localized by tracing the course of a particular vessel; these defects correspond anatomically to the normal distribution of the retinal vasculature (Figs 1.26 and 1.27). Thus retinal vascular filling defects result from a variety of disease processes, but most are commonly associated with atherosclerosis and diabetes.

Fig. 1.26 Branch retinal artery occlusion. (A) Color photograph showing areas of retinal whitening inferior to the macula. An intra-arterial embolus is seen proximal to the whitened retina. (B) Earliest arterial filling. An area of hypofluorescence is located inferior to the macula. A small retinal artery feeding this area is occluded at the site of the embolus previously identified on Figure 1.26A. (C) Midarterial venous filling with a small area of intra-arterial hyperfluorescence distal to the site of embolic obstruction. (D) Late arteriovenous phase of fluorescein angiogram shows that the occluded artery is still mostly hypofluorescent. (E) Portions of the area did fill with fluorescein due to retrograde filling from surrounding areas. There is some mild staining of the occluded retinal artery.

Fig. 1.27 Retinal branch-vein occlusion. (A) Color photograph of the right macula and disc. There are areas of retinal hemorrhage, retinal whitening, and cotton-wool spots. (B) The fluorescein angiogram of the right disc and macula shows normal fluorescence of the superior portion of the macula. The inferior portion shows substantial hypofluorescence due to retinal capillary nonperfusion. The very bright hyperfluorescent areas are due to neovascularization. Comment: This patient had a very ischemic inferotemporal branch retinal vein occlusion of the right eye. This was a severe occlusion, as evidenced by closure of large areas of the capillary bed. The hypofluorescence was caused not only by vascular filling defect but also by the nonperfused retina, which becomes partially opaque and caused hypofluorescence of the choroid. (In other words, there was blockage of choroidal fluorescence by the opaque retina, which was caused by the retinal capillary nonperfusion.)

Vascular filling defects of the disc

Vascular filling defects of the disc occur because of the failure of the capillaries of the optic nerve head to fill. This failure can be caused by: (1) congenital absence of disc tissue, as in an optic pit or optic nerve head coloboma (Fig. 1.28); (2) atrophy of the disc tissue and its vasculature, as in optic atrophy; or (3) vascular occlusion, as in an ischemic optic neuropathy.^19,20 Each condition is characterized by early hypofluorescence caused by nonfilling and late hyperfluorescence resulting from staining of the involved tissue.

Fig. 1.28 Optic pit and sensory macula detachment. (A) Color photograph of left macula. Note the dark area of the optic pit (arrows). Cystic edema is present in the macula secondary to the macular schisis detachment from the optic pit. (B) Early arteriovenous phase of fluorescein angiogram shows hypofluorescence of the disc in the area of the pit due to absence of tissue and vessels. (C) In the late arteriovenous phase fluorescein angiogram, the hypofluorescent area of the pit is evident.

Choroidal vascular filling defect

The normal choroidal vasculature is usually difficult to document with fluorescein angiography because of the pigment epithelial barrier. If chronic choroidal vascular filling defects exist, the pigment epithelium is often secondarily depigmented or atrophied. In these cases the hypofluorescence caused by a vascular filling abnormality of the choroid and choriocapillaris can be documented angiographically.

When choroidal vessels do not fill, dark patches of hypofluorescence beneath the retina appear early in the angiogram. The distribution and morphology of the hypofluorescence vary according to the disease process. Because the choroidal circulation is completely separate from the retinal circulation, choroidal vascular filling defects do not correlate with the retinal vascular distribution. If the choriocapillaris is absent and the large choroidal vessels are still present, the choroidal and retinal vessels fluoresce, but hypofluorescent gaps appear because of the loss of the diffuse “ground-glass” fluorescence from the choriocapillaris (Fig. 1.29). When the choroidal vasculature does not fill, as in total occlusion or in atrophy, hypofluorescence occurs early in the angiogram. The hypofluorescence remains throughout the late stages of the procedure, although leakage from surrounding areas of normal choriocapillaris extends into the occluded area. When sufficient leakage occurs, the sclera retains fluorescein (stains) late in the angiogram. When the involved area is large and the leakage is minimal, the hypofluorescence remains throughout the later stages.

Fig. 1.29 Choroideremia: total loss of retinal pigment epithelium (RPE) and choriocapillaris with much of the large choroidal vasculature remaining. (A) Color photograph of the left disc and macula. The large choroidal vasculature can be seen as pale, irregular lines. Dark patches of pigment are located in the macula and around the disc. (B) Late venous phase of fluorescein angiogram of the left disc and macula. The large choroidal vessels can be seen filling, as can the retinal arteries. The choriocapillaris is not seen. (C) Recirculation phase of fluorescein angiogram. The large choroidal vessels and retinal vessels can be seen, but the choriocapillaris (usually seen as ground-glass fluorescence) is not seen except in the far edges of the view. Comment: This patient had a total loss of RPE and choriocapillaris in most areas of the fundus. The ground-glass choroidal fluorescence was absent from most areas. The large choroidal vessels could be seen. The large choroidal vessels do not leak fluorescein, and therefore the sclera did not stain in these areas. The RPE and choriocapillaris were partially intact in a few areas. These can be seen at the far extremes, where there is some mild ground-glass appearance.

A normal physiologic condition exists in many patients in which the choroid fills in a patchy manner. Areas adjacent to the foci that are filling show early hypofluorescence but eventually fill normally, usually 2–5 seconds later. This has been termed patchy choroidal filling, and it is the most common form of choroidal vascular filling defect. This form of filling follows a pattern in which the short posterior ciliary arteries enter the eye perpendicularly through the sclera. These vessels then feed the choriocapillaris lobules.

The prechoriocapillaris arterioles and lobules are end, or terminal, vessels demonstrating no anastomoses with adjacent choriocapillaris arterioles or lobules. Each choriocapillaris lobule is connected to adjacent lobules on the venous, or emptying, side of the circulation. Fluorescence in each choriocapillaris segment or lobule is in the form of a round, irregular, or hexagonal patch. When some of the channels fill late, a heterogeneous filling pattern results. The choriocapillaris fills most areas, whereas dark hypofluorescent patches are present in other areas. These dark areas are lobules from separate end channels that are not filled simultaneously with adjacent choriocapillaris lobules. They are filled in a delayed fashion by the single feeder choroidal arteriole.

In general, vascular filling defects of the choroid are caused by obstructive disorders or absence of tissue with the following fluorescein angiographic characteristics: (1) normal retinal vascular flow; (2) depigmentation of the pigment epithelium; (3) reduction of choroidal blood flow; and (4) hypofluorescence in the early phases of angiography caused by loss of the normal ground-glass choriocapillaris fluorescence. In some conditions the large choroidal vessels are also absent, resulting in total early hypofluorescence in the affected area, with scleral staining only on the circumference of the lesion because of the adjacent patent choriocapillaris. Choroidal vascular defects result from a variety of disease processes (Figs 1.30 and 1.31).

Fig. 1.30 Choroidal atrophy, with some remaining islands of choriocapillaris, due to choroideremia. (A) Schematic of retina shows loss of pigment epithelium and choriocapillaris and some of the outer retina (especially photoreceptors). (B) Color photograph of left superior retina showing areas of severe atrophy and more intact areas of retinal pigment epithelium (RPE) peripherally. Arrows delineate margins between normal RPE and RPE atrophy producing window defect. (C) The arteriovenous phase of fluorescein angiogram shows normal fluorescence of the retinal arteries. The large choroidal vessels can be seen temporally on the right side of the photograph. The ground-glass fluorescence of the choriocapillaris can be seen more peripherally on the left side of the angiogram, where the RPE and choriocapillaris are more intact. Comment: This patient had severe atrophy of the RPE and choriocapillaris. Large choroidal vessels could be seen causing hypofluorescence in relationship to absence of ground-glass choroidal fluorescence. Some areas of choriocapillaris remained and showed normal hyperfluorescence (perhaps increased hyperfluorescence caused by loss of overlying RPE).

Fig. 1.31 Choroidal hypoperfusion caused by photodynamic therapy with verteporfin. (A) Left macula. Color photograph shows widespread retinal pigment epithelial alterations and drusen secondary to an occult choroidal neovascular membrane secondary to age-related macular degeneration. This treatment was considered standard therapy prior to the advent of intraocular antivascular endothelial growth factor medications. (B) The late arteriovenous phase of the fluorescein angiogram of the left macula shows hypofluorescence of the macula and a large area temporally. Larger choroidal vessels are perfused. The macular hypofluorescence corresponds to the laser treated area. The large area of hypofluorescence temporally represents an area of choroidal nonperfusion caused by selective choriocapillaris occlusion from photodynamic therapy. (C) Later phase of the fluorescein angiogram shows continued hypofluorescence of the area temporal to the macula despite relative restoration of perfusion to choriocapillaris temporal to macula.

Hyperfluorescence

Hyperfluorescence is any abnormally light area on the positive print of an angiogram, that is, an area showing fluorescence in excess of what would be expected on a normal angiogram. There are four possible causes of abnormal hyperfluorescence: (1) preinjection fluorescence; (2) transmitted fluorescence; (3) abnormal vessels; and (4) leakage. The appearance of fluorescence depends in part on the relationship of its appearance to the timing of the fluorescein injection.

Preinjection fluorescence is hyperfluorescence that can be seen before fluorescein is injected and is caused by structures that naturally fluoresce (autofluorescence) or by poorly matched filters (pseudofluorescence).

Transmitted fluorescence and abnormal vascular fluorescence occur in the early, or vascular, stage of the angiogram, when fluorescein fills patent blood vessels. Transmitted fluorescence appears when fluorescein fills the normal choriocapillaris, but it is more noticeable when there is reduced pigment in the pigment epithelium or loss of retinal pigment epithelium. This is designated pigment epithelial window defect.

When abnormal retinal, disc, or choroidal vessels are present and fill with fluorescein, hyperfluorescence occurs. This type of hyperfluorescence, abnormal vascular fluorescence, is also seen in the early, or vascular, phase of the angiography.

Hyperfluorescence caused by leakage is seen predominantly in the later, or extravascular, phase of angiography. In this phase, fluorescein has emptied from normal and abnormal vessels. Any significant fluorescein that remains in the eye is fluorescein that has escaped or leaked from vascular or tissue barriers and is thus extravascular.

Therefore, to ascertain the type of hyperfluorescence, one must determine the time at which the hyperfluorescence appears in relation to when the fluorescein was injected. Once the hyperfluorescence is determined to be caused by preinjection fluorescence, transmitted fluorescence, the presence of abnormal vessels, or by leakage, the next step is to determine the anatomic location of the hyperfluorescence. Abnormal blood vessels may come from the retina and disc or from the choroid. Leakage can occur in the vitreous, disc, retina, or choroid.

Autofluorescence

Autofluorescence is the emission of fluorescent light from ocular structures in the absence of sodium fluorescein. Conditions that cause autofluorescence are optic nerve head drusen and astrocytic hamartoma (Fig. 1.32).

Fig. 1.32 Autofluorescence of optic nerve drusen. (A) Right disc and macula show a blurred disc margin with nonhyperemic vessels. Blurring of the central optic nerve is consistent with disc edema noted on stereophotos. (B) Preinjection or “control” photos are performed with filters in place, prior to any injection of fluorescein. This allowed for the identification of optic nerve drusen, which autofluoresce.

Pseudofluorescence occurs when the blue exciter and green barrier filters overlap. The blue filter overlaps into the green range, allowing the passage of green light, or the green barrier filter overlaps into the blue range, allowing the passage of blue light (Fig. 1.2). The overlapping light passes through the system, reflects off highly reflective surfaces (light-colored or white structures), and stimulates the film. This reflected nonfluorescent light is called pseudofluorescence.

Conditions that tend to produce pseudofluorescence include any light-colored or white (reflective) fundus change (e.g., sclera, exudate, scar tissue, myelinated nerve fibers, foreign body).

Currently, fluorescein angiographic filters are usually very well matched; overlap is minimal, so pseudofluorescence is faint and rarely a major problem. However, filters do tend to get thin with time. The frequent flashes of light from the fundus camera wear them down, and most filter pairs eventually allow pseudofluorescence. Therefore, depending on frequency of use, fluorescein filters must be changed occasionally. Our experience indicates that change is required approximately every 5 years.

Transmitted fluorescence (pigment epithelial window defect)

This fluorescence is an accentuation of the visibility of normal choroidal fluorescence. Transmitted fluorescence occurs when fluorescence from the choroidal vasculature appears to be increased because of the absence of pigment in the pigment epithelium, which normally forms a visual barrier to choroidal fluorescence. The major cause of pigment epithelial window defect is atrophy of the pigment epithelium (Figs 1.33–1.36).

Fig. 1.33 Pigment epithelial window defect. This schematic of the retina shows that the pigment epithelium in the center of the section is less pigmented than the normal pigment epithelium. This allows the normal choroidal and choriocapillaris fluorescence to show through; that is, this pathologic condition would create a typical pigment epithelial window defect.

Fig. 1.34 An eye with drusen demonstrating pigment epithelial window defects. (A) Color photograph: right macula shows multiple drusen temporally. (B) Late arteriovenous phase of fluorescein angiogram shows marked hyperfluorescence in the areas of the drusen. (C) Late recirculation phase of fluorescein angiogram shows fading of fluorescence. Comment: Note the degree of fluorescence of the entire fundus vasculature. This is typical of a pigment epithelial window defect, which is a type of vascular fluorescence. The drusen allow a better view to the choriocapillaris because of the thinning of the pigment epithelium overlying them.

Fig. 1.35 Pigment epithelial window defect: choroidal folds. (A) Montage color photograph of right disc and macula. Note the pale lines (choroidal folds) scattered throughout the posterior pole. (B) Arteriovenous phase of fluorescein angiogram of the disc and macula. Hyperfluorescent lines correspond to the folds, and adjacent hypofluorescent lines are present throughout the macula and surrounding the disc. Comment: This patient had pigment epithelial folds caused by prolonged hypotony from a filtering bleb. The hyperfluorescent lines are thought to be the hills of the folds, in the apices of which the pigment epithelium is thinned, allowing hyperfluorescence in the early phases of the fluorescein angiogram (pigment epithelial window defect). The dark lines are thought to be the valleys of the folds, with an increase in pigmentation causing blockage of choroidal fluorescence. The later phases of fluorescein angiograms often show fading of fluorescence. Choroidal folds represent a type of pigment epithelial window defect with early vascular fluorescence and late fading of fluorescence.

Fig. 1.36 Pigment epithelial window defect: macular hole. (A) Schematic drawing of macula showing loss of entire central foveal tissue. (B) Color photograph of the left macula. This patient has a macular hole. Note a corona of lighter detached, swollen retinal tissue surrounding the foveal center where the hole is located. (C) Late phase of fluorescein angiogram shows hyperfluorescence within the macular hole. (D) Later phase of the fluorescein angiogram shows some fading of the hyperfluorescence within the macular hole. Comment: The choriocapillaris was intact. Therefore the angiogram showed normal fluorescence of the choriocapillaris (early hyperfluorescence within the center of the fovea) and fading in the late phase of the angiogram.

When the pigment epithelium is dense, choroidal fluorescence is not clearly visible because the pigment blocks the view of the choroid and acts as a barrier to fluorescein. The density of the pigment determines the degree to which transmission of the normal choroidal fluorescence is blocked. The visibility of choroidal fluorescence is inversely proportional to the concentration of pigment in the pigment epithelium. If the pigment epithelium contains less than the normal amount of pigment or is defective, the choriocapillaris appears to fluoresce more brightly. The presence of hyperfluorescence caused by a defect in the pigment epithelium depends on the state of both the pigment epithelium and the choriocapillaris. The choriocapillaris must be intact for a depigmented area of the pigment epithelium to be apparent. If the choriocapillaris does not fill, a depigmented area of the pigment epithelium does not fluoresce.

Transmitted fluorescence has the following four basic characteristics:

1. It appears early in angiography, coincidental with choroidal filling.

2. It increases in intensity as dye concentration increases in the choroid.

3. It does not increase in size or shape during the later phases of angiography.

4. It tends to fade and sometimes disappear as the choroid empties of dye at the end of angiography.

In short, transmitted fluorescence appears, peaks early, and fades late without changing size or shape, as would any normal vascular fluorescence. When pigment epithelial depigmentation is extensive, late fluorescein staining of the choroid and sclera may be visible, although it is less intense than the fluorescence of the window defect.

Abnormal retinal and disc vessels

Abnormal vascular fluorescence occurs when abnormal vessels are present. Such pathologic vessels may be in the retina, on the disc, or at the level of the choroid. Normal and abnormal retinal and disc vessels are clearly visible on the angiogram because no barrier obscures them from view. Gross abnormalities of the retinal and disc vasculature and subtle microvascular changes that cannot be appreciated adequately by ophthalmoscopic examination will be well defined and easily distinguished by fluorescein angiography. These changes in the retinal vasculature can be classified into six morphologic categories: (1) tortuosity and dilation (Figs 1.37 and 1.38); (2) telangiectasis (Figs 1.39 and 1.40); (3) neovascularization (Fig. 1.41); (4) anastomosis (Fig. 1.38); (5) aneurysms (Figs 1.38 and 1.39); and (6) tumor vessels (Figs 1.42 and 1.43).

Fig. 1.37 Abnormal retinal vessels, tortuosity, and dilation: internal limiting membrane contraction. (A) Color photograph of right macula shows a pale membrane overlying the right macula, producing contraction of the retina and tortuosity of the retinal vessels. (B) Arteriovenous phase of fluorescein angiogram shows marked irregularity and tortuosity of the retinal vessels in association with the preretinal membrane (macular pucker). (C) Late phase of fluorescein angiogram shows a small amount of vascular leakage due to contraction of the membrane and pulling on the retinal vessels. Comment: This is tortuosity and dilation, a type of abnormal retinal vascular fluorescence. It is caused by the mechanical traction of an epiretinal membrane.

Fig. 1.38 Abnormal retinal vascular fluorescence: retinal vascular microaneurysms, telangiectasis, and anastomoses. (A) Color photography of right eye shows numerous telangiectatic retinal vessels due to a superotemporal branch-vein occlusion. (B) Arteriovenous-phase fluorescein angiogram shows multiple areas of smaller and larger microaneurysms and telangiectasis. Several small venous–venous anastomoses can be seen just temporal to the macula. The venous system of the occluded area has collateralized with patent vessels in uninvolved areas.

Fig. 1.39 Retinal telangiectasis and microaneurysms secondary to diabetic retinopathy. (A) Color photography of right macula showing retinal exudate, retinal striae, and irregularly dilated retinal vessels (telangiectasis). (B) Arteriovenous-phase fluorescein angiogram shows extensive hyperfluorescence from the numerous microaneurysms, and telangiectatic retinal vessels. (C) Later arteriovenous-phase fluorescein angiogram of right macula showing leakage from many of these vessels. (D) Late-phase fluorescein angiogram of right macula shows multiple circular areas of hyperfluorescence due to accumulation of dye in extensive cystoid spaces. Comment: This patient had significant retinal microvascular changes due to diabetic retinopathy.

Fig. 1.40 Abnormal retinal vessels: telangiectasis. (A) Color montage photograph demonstrating severe areas of exudation, as well as dilated and telangiectatic vessels. The retina is very edematous. (B) Arteriovenous phase of fluorescein angiogram shows marked irregularity of the retinal vasculature. There are areas of capillary nonperfusion, telangiectasis, and tortuosity. Comment: This patient had Coats disease with a markedly abnormal retinal capillary bed, including telangiectasis and dilated vessels.

Fig. 1.41 Abnormal retinal vessels: retinal neovascularization due to proliferative diabetic retinopathy. (A) Montage color photograph of the posterior pole of the right eye. Extensive irregular tortuous vessels extend from the optic nerve along the vascular arcades and nasally. These vessels lie on the surface of the retina. (B) Later arteriovenous phase of fluorescein angiogram montage shows increasing hyperfluorescence of the retinal neovascularization. Comment: This patient had severe proliferative diabetic retinopathy with extensive neovascularization of the right disc. The vessels fluoresced early (vascular fluorescence) and leaked late. This is very typical of retinal or disc neovascularization.

Fig. 1.42 Abnormal retinal vessels: tumor–retinal angioma as part of von Hippel’s disease. (A) Color photograph of right macula and disc shows exudate temporal and inferior to the disc. A very vascular, slightly elevated mass was noted on the temporal border of the disc. Ophthalmoscopy showed that it has a reddish appearance. A large full-thickness macular hole is also observed. (B) Early arterial phase of the fluorescein angiogram shows marked fluorescence of the mass. (C) Midarteriovenous phase of the fluorescein angiogram shows an increased fluorescence of the mass. (D) Late phase of the fluorescein angiogram shows leakage of fluorescein within the mass. Comment: This patient had a peripapillary retinal angioma. It was very vascular and showed early fluorescence and extensive late leakage.

Fig. 1.43 Arteriovenous malformation: Wyburn–Mason type. (A) Color montage photograph of right macula and temporal retina showing enlarged, dilated retinal artery, with direct connection to an engorged draining vein. There is no intervening capillary bed. (B) Fluorescein angiogram showing marked hyperfluorescence of the abnormal, dilated retinal artery and vein. Two smaller arteriovenous malformations appear to be present, one just above the macula, and the other just below.

These aforementioned changes can be viewed in the early (vascular) phases of angiography. Later, as the vessels empty, some of these vascular abnormalities leak fluorescein, whereas others do not.

Vascular abnormalities of the retina and disc are readily apparent on the fluorescein angiogram. The changes are characterized by early vascular-appearing hyperfluorescence. Each of the six morphologic types indicates specific disease processes that aid the clinician in making a diagnosis, determining the degree of the distinct pathologic process, and understanding the pathophysiology of retinal vascular disease.

Abnormal choroidal vessels

Abnormal vessels that may be present under the retina and originate from the choroid are subretinal neovascularization and vessels within a tumor. When subretinal neovascularization is present, the early angiogram often shows a lacy, irregular, and nodular hyperfluorescence (Figs 1.44 and 1.45). With a choroidal tumor, the abnormal hyperfluorescence is a similar, early vascular-type fluorescence, although it may be coarser, as seen in choroidal hemangioma (Fig. 1.46) and malignant melanoma (Fig. 1.47).

Fig. 1.44 Abnormal choroidal vessels: subretinal neovascularization. (A) Schematic view of the retina shows a small break in Bruch’s membrane, with a fine proliferation of capillaries through the break dissecting under and lifting up the pigment epithelium. There is a shallow sensory retinal detachment. (B) Color photograph of the left macula. There is a dirty-gray membrane involving the central macula. Note the small area of subretinal hemorrhage. There is a shallow sensory retinal detachment. (C) The arteriovenous phase of fluorescein angiogram shows fine, lacy, irregular hyperfluorescence corresponding to a small, fine patch of subretinal neovascularization. (D) Late phase of fluorescein angiogram shows leakage of these vessels into the subpigment epithelial and subretinal spaces. Comment: This patient had a small patch of subretinal neovascularization involving the central fovea. The angiogram shows typical, early vascular fluorescence (in a nodular, irregular, lacelike fashion) and late hyperfluorescent leakage.

Fig. 1.45 Abnormal choroidal vessels: subretinal neovascularization. (A) Schematic drawing of retina shows vascular proliferation from the choriocapillaris dissecting under the pigment epithelium, with associated fibrous tissue. The pigment epithelium has become thinned and the sensory retina detached. The outer plexiform layer of the sensory retina shows cystic spaces. (B) Red-free photograph of left macula shows some hemorrhage and exudate. On the color photograph and slit-lamp biomicroscopy, a dirty-gray membrane was noted in the inferotemporal portion of the macula. This is seen as a slightly pale lesion in the inferotemporal macula. (C) Early arteriovenous phase of fluorescein angiogram shows a lacy, irregular, nodular area of hyperfluorescence in the inferotemporal macula. This is a flat patch of vessels that has proliferated from the choriocapillaris under the pigment epithelium. (D) Late phase of the fluorescein angiogram shows leakage from the patch of subretinal neovascularization. Most of the fluorescence is pooling of fluorescein under the sensory retinal detachment, although there is some cystic change in the fovea. Comment: This patient had a patch of subretinal neovascularization that was nearly 4 disc diameters in size. It fluoresced early with the vascular phase of the angiogram (typical for subretinal neovascularization) and leaked late. Actually, “subretinal neovascularization” is a misnomer because the new vessels are initially located in the subpigment epithelial space.

Fig. 1.46 Abnormal choroidal vessels in a patient with choroidal hemangioma. (A) Color photos of left macula and disc with elevated choroidal hemangioma. (B) Arteriovenous phase of the fluorescein angiogram shows prominent hyperfluorescence in this area demonstrating the tumor vessels. (C) Late phase of the fluorescein angiogram shows marked leakage in this area. Comment: This patient had a choroidal hemangioma, which is a very vascularized choroidal mass. The vascularity in this mass causes the marked hyperfluorescence and leakage.

Fig. 1.47 Abnormal choroidal vascular fluorescence due to malignant melanoma. (A) Color photograph of left eye. Note the darkly pigmented mass nasal to the optic nerve. There is some orange lipofuscin pigment overlying the surface of this as well. (B) Arteriovenous phase of fluorescein angiogram of the mass shows hyperfluorescence over the surface of the tumor. This patient also had some macular drusen, which show some early hyperfluorescence in the macula. (C) Late phase of the fluorescein angiogram shows leakage from the mass. There are multiple “hot spots” overlying the tumor. Comment: This patient had a choroidal malignant melanoma. This was a medium-sized tumor that showed the typical early fluorescence that is seen in a medium-sized melanoma.

Leak

Fluorescence of the retinal and choroidal vessels begins to diminish about 40–60 seconds after injection. Fluorescein empties almost completely from the retinal and choroidal vasculature about 10–15 minutes after injection. Any fluorescence that remains in the fundus after the retinal and choroidal vessels have emptied of fluorescein is extravascular fluorescence and represents leakage.

Four types of late extravascular hyperfluorescent leakage occur in the normal eye: (1) fluorescence of the disc margins from the surrounding choriocapillaris; (2) fluorescence of the lamina cribrosa; (3) fluorescence of the sclera at the disc margin if the retinal pigment epithelium terminates away from the disc, as in an optic crescent; and (4) fluorescence of the sclera when the pigment epithelium is lightly pigmented. These are the only forms of late hyperfluorescence or leakage that can be considered “normal.” Any other hyperfluorescence observed 15 minutes after the fluorescein injection represents extravascular fluorescein and is referred to as leakage.

Either or both of the two vascular systems of the fundus can produce abnormal late hyperfluorescence (leakage) if defects are present in their respective barriers to fluorescein. The barrier to fluorescein leakage from the retinal vessels is the retinal vascular endothelium. The barrier to leakage from the choroidal circulation is the pigment epithelium. An abnormality of the retinal vascular endothelium can result in permeability to fluorescein and leakage of fluorescein into the retinal tissue. Similarly, an abnormality of the pigment epithelium can result in permeability to fluorescein, and fluorescein will leak from the choroidal tissue through the pigment epithelium. Abnormal late hyperfluorescence of the choroid, however, can occur without damage to the pigment epithelium, as in cellular infiltrates of the choroid that occur in choroidal inflammation or tumor.

There are two other types of late abnormal fluorescence: one occurs when fluorescein enters the vitreous, and the other when fluorescein leaks into the optic nerve head.

Vitreous leak

Leakage of fluorescein into the vitreous creates a diffuse, white haze in the late phase of the fluorescein angiogram. In some instances the haze is generalized and evenly dispersed, and in other cases the white haze is localized.

Leakage of fluorescein into the vitreous is due to three major causes: (l) neovascularization growing from the retinal vessels on to the surface of the retina or disc or into the vitreous cavity; (2) intraocular inflammation; and (3) intraocular tumors.

Vitreous hyperfluorescence secondary to retinal neovascularization is usually localized and appears as a cotton-ball type of fluorescence surrounding the neovascularization (Fig. 1.41B). The vitreous fluorescence secondary to intraocular inflammation is often generalized, giving a diffuse, white haze to the vitreous because of generalized leakage of fluorescein from the iris and ciliary body. The vitreous fluorescence secondary to tumors is most often localized over the tumor.

Disc leak

The optic nerve head normally has some fluorescein leakage (late hyperfluorescence) as a result of staining of the lamina cribrosa and the surrounding margins of the disc (from the normally leaking peripapillary choriocapillaries). The difference between normal and abnormal leakage at the disc may be subtle.

Papilledema and optic disc edema

Papilledema is swelling of the optic nerve head as a result of increased intracranial pressure. Edema of the optic disc is defined as swelling of the optic nerve head secondary to local or systemic causes (Fig. 1.48). The angiogram is similar in each case, demonstrating leakage associated with swelling of the optic nerve head. In the early phases of the angiogram, dilation of the capillaries on the optic nerve head may be seen; in the late angiogram, the dilated vessels leak, resulting in a fuzzy fluorescence of the disc margin.

Fig. 1.48 Disc leakage. (A) Color photograph of right optic nerve. Note the dilation of the disc capillaries. (B) The arteriovenous-phase angiogram of the right disc and macula shows the hyperfluorescence due to these dilated disc capillaries. (C) The late phase of the angiogram shows significant leakage from these dilated optic disc capillaries. Comment: This patient had a papillopathy related to diabetes. This produced significant dilation of the disc capillaries. The leakage from this abnormal disc is quite obvious.

Retinal leak

In the late stages of the normal angiogram, the retinal vessels have emptied of fluorescein and the retina is dark. Any late retinal hyperfluorescence is abnormal and indicates leakage of retinal vessels. When the leakage is severe, the extracellular fluid may flow into cystic pockets, and the angiogram shows fluorescence of the cystic spaces. Fluorescein flows out of the patent retinal vessels to lie in pools in the cystoid spaces or stains the edematous (noncystic) retinal tissue. Cystoid retinal edema is apparent as the fluorescein pools in small loculated pockets. In the macula, cystoid edema takes on a stellate appearance (Fig. 1.49); elsewhere in the retina, it has a honeycombed appearance (Fig. 1.50). Fluorescent staining of noncystoid edema is diffuse, irregular, and not confined to well-demarcated spaces (Figs 1.51 and 1.52).

Fig. 1.49 Retinal leak: cystoid macular edema. (A) Schematic drawing of the macula shows large cystic spaces in the outer plexiform layer. There are some cystic spaces in the inner nuclear layer. (B) Color photograph of left macula. Careful inspection of the retina is often necessary on biomicroscopy to detect intraretinal cystoid. (C) Arteriovenous phase of fluorescein angiogram shows some dilation of the fine capillary network around the fovea. (D) Late phase of fluorescein angiogram shows hyperfluorescence from the accumulation of dye filling the cystic spaces. Note the stellate appearance of the cystoid macular edema. Comment: This patient had late hyperfluorescence (i.e., leakage) into the retina that was severe enough to create cystic spaces. This is a typical example of cystoid macular edema.

Fig. 1.50 Retinal leakage: cystoid retinal edema. (A) Color photograph of the right macula. Large central cystoid cavity is seen corresponding to fovea. (B) Arteriovenous phase of fluorescein angiogram shows well-defined telangiectatic retinal vessels. (C) Late phase of fluorescein angiogram shows leakage from these vessels. In the center of the macula, the leakage is in stellate cystic pockets, and just outside the macula, temporally, the leakage has taken a honeycomb form. Comment: This patient had leakage of telangiectatic vessels into the retina, and the leakage formed cystoid spaces. Cystoid edema in the center of the macula takes on a stellate form because of the oblique nature of the outer plexiform layer. The cystic spaces take on a honeycomb form in nonmacular areas of the retina because of the perpendicular nature of the fibers of the outer plexiform layer.

Fig. 1.51 Retinal leakage, severe noncystoid edema. Branch-vein occlusion. (A) Color photograph of right macula shows multiple retinal hemorrhages inferotemporally due to a retinal branch-vein occlusion. (B) Arteriovenous phase of fluorescein angiogram shows the vascular abnormalities associated with the branch-vein occlusion. Hypofluorescence corresponds to areas previously treated with grid pattern laser photocoagulation. (C) Late phase of fluorescein angiogram shows diffuse leakage of the fluorescein dye. Comment: This patient had generalized leakage of the retinal vascular bed in the distribution of the blocked branch vein. The leakage was not yet severe enough, however, to form clearly defined cystic spaces. Late hyperfluorescence indicates leakage, and this fluorescence is located in the retina; thus this was retinal edema.

Fig. 1.52 Late hyperfluorescence, retinal leakage: severe epiretinal membrane contraction. (A) Color photograph of right macula showing thick epiretinal membrane overlying the macula and producing severe traction and contraction of the retina and vessels. (B) Arteriovenous phase of fluorescein angiogram shows that the retinal vasculature is tortuous and irregular. (C) Late arteriovenous phase of fluorescein angiogram shows leakage from the retinal vessels. Comment: The marked preretinal membrane caused sufficient traction on the retina, resulting in marked retinal vascular leakage.

The amount of fluorescein leakage depends on the dysfunction of the retinal vascular endothelium (Fig. 1.52). When leakage is not pronounced, the cystoid spaces fill slowly and become visible only late in angiography. When this occurs, the area of cystoid retinal edema may be somewhat hypofluorescent early in the angiogram because the fluid in these spaces acts as a barrier and blocks the underlying choroidal fluorescence. When there is heavy fluorescein leakage, the cystoid spaces fill rapidly, in some cases within a minute after injection. The large confluent cysts seen with severe cystoid macular edema may fill late in the angiogram. The large retinal vessels can also leak. This is called perivascular staining and is seen in three distinct situations: inflammation (indicating a perivasculitis), traction (severe pulling on a large retinal vessel, Fig. 1.52), and occlusion. When a large retinal vessel leak is partially occluded, or when it traverses an area of occlusion (and capillary nonperfusion), it will leak (Fig. 1.53).

Fig. 1.53 Retinal leakage: perivascular staining. (A) In this late arteriovenous-phase fluorescein angiogram, note the beading of the large retinal veins. There is also associated leakage from these vessels. (B) Later phase of fluorescein angiogram shows perivascular staining (leakage) from the large retinal vessels that are traversing large zones of capillary nonperfusion. Comment: Typically, when a large retinal vessel (artery or vein) is perfused but traverses an area of capillary nonperfusion, ischemic retinal factors will act adversely on the endothelium of the large vessel and cause it to leak. This is called perivascular staining. Perivascular staining also occurs with traction or inflammation.

Choroidal leak

Late hyperfluorescence under the retina can be classified as either pooling or staining (Fig. 1.54). Pooling is defined as leakage of fluorescein into a distinct anatomic space; staining is leakage of fluorescein diffused into tissue.

Fig. 1.54 Late hyperfluorescence, subretinal pooling: central serous chorioretinopathy. (A) Schematic drawing of retina shows a sensory retinal detachment. There is a break in the pigment epithelium. Fluorescein flows from the choriocapillaris through Bruch’s membrane, through the break in the pigment epithelium, and into the subretinal space, under the detached retina. (B) Color photograph of left macula shows a shallow sensory detachment (arrows). Just superonasal to the fovea is a small white area with a gray center. The fluorescein angiogram will reveal that this is the area of the leak. (C) Arteriovenous phase of fluorescein angiogram shows a hyperfluorescent spot that was seen on stereoangiography to be leakage of fluorescein coming from the pigment epithelium. (D) Late phase of fluorescein angiogram shows that the spot of pigment epithelial leakage has enlarged and become fuzzy. This is the release of fluorescein molecules into the fluid under the detached sensory retina. Comment: This patient had central serous chorioretinopathy. There was a break in the pigment epithelium that allowed leakage of fluorescein through it and into the subretinal space. Late hyperfluorescence means leakage, and in this case, there is pooling of fluorescein under the detached retina.

Fluorescein pools in the spaces created by detachment of the sensory retina from the pigment epithelium or in the space created by detachment of the pigment epithelium from Bruch’s membrane. The posterior layer of the sensory retina is made up of rods and cones that are loosely attached to the pigment epithelium. When a sensory retinal detachment occurs, the detached segment separates with little force, forming a very gradual angle at the point of attachment to the pigment epithelium. Because of this narrow angle, the exact limits of a sensory retinal detachment are difficult to locate ophthalmoscopically or by slit-lamp biomicroscopy.

Depending on the specific disease, the late angiogram may or may not portray the full fluorescent filling of the subretinal fluid. For example, in central serous chorioretinopathy the leakage is gradual, and fluorescence of the subsensory retinal fluid will not be complete. In other conditions, such as subretinal neovascularization, fluorescein leakage is profuse, and the subsensory fluid often completely fluoresces (Fig. 1.55).

Fig. 1.55 Late hyperfluorescence, leakage, and pooling under the sensory retina caused by subretinal neovascularization, resulting in a sensory retinal detachment. (A) Schematic of the retina showing that it has detached (photoreceptors are separated from pigment epithelium). Vessels have proliferated from the choriocapillaris through Bruch’s membrane. There is a fibrovascular scar involving the pigment epithelium. The sensory retina is detached. (B) Color photograph of left macula shows a pale gray lesion in the inferior portion of the macula with some associated hemorrhage. (C) Arteriovenous phase of fluorescein angiogram shows a patch of subretinal neovascularization inferior to the fovea; this is evidenced by the lacy, irregular hyperfluorescence in this area. (D) Late phase of fluorescein angiogram shows fuzzy fluorescence. There is pooling of fluorescein under the detached retina and some staining of the fibrous tissue associated with the subretinal neovascularization. Comment: This patient had a patch of subretinal neovascularization with a great deal of leakage, causing a sensory detachment. The early-phase angiogram showed the vascular nature of the lesion, and the late-phase angiogram showed the leakage and pooling in the subretinal space.

In contrast to the attachment of the sensory retina, the basement membrane of the pigment epithelium adheres firmly to the collagenous fibers of Bruch’s membrane. The firm adhesion and wide angle of detachment make it easy to discern a pigment epithelial detachment ophthalmoscopically. Occasionally a light-orange ring appears around the periphery of a pigment epithelial detachment, further facilitating identification (Fig. 1.56).

Fig. 1.56 Late hyperfluorescent pooling under the retinal pigment epithelium (pigment epithelial detachment). (A) Schematic diagram illustrating detachment and elevation of the pigment epithelium; the pigment epithelium is separated from Bruch’s membrane. Because the attachment of the pigment epithelium to Bruch’s membrane is quite firm, the angle of detachment is quite large. (B) Color photography of right macula shows a round detachment of the pigment epithelium. (C) Early arteriovenous phase of fluorescein angiogram shows early fluorescence from the area of detachment pigment epithelium. (D) Late-phase angiogram of right macula shows well-demarcated hyperfluorescent pooling of fluorescein under the detached pigment epithelium. Comment: Fluorescein flows freely through Bruch’s membrane and stops at the pigment epithelium. When the pigment epithelium is detached, the fluorescein flows right through Bruch’s membrane into the space made by the detached pigment epithelium. Therefore a pigment epithelial detachment fluoresces evenly and slowly (like a light bulb on a rheostat) and shows intense hyperfluorescent pooling that is well demarcated (indicating its well-defined angle of attachment) late in the angiogram.

The differences in the adherence and the angle of detachment between a sensory retinal detachment and a pigment epithelial detachment result in specific differences in fluorescent pooling patterns. The hyperfluorescent pooling of a sensory retinal detachment tends to fade gradually toward the site where the sensory retina is attached. This makes fluorescein angiographic determination of the extent of a sensory retinal detachment difficult. In contrast, the hyperfluorescent pooling under a pigment epithelial detachment extends to the edges of the detachment, making the entire detachment and its margins hyperfluorescent and clearly discernible.

Pooling of fluorescein under a sensory retinal detachment in central serous retinopathy takes place slowly, since the dye passes through one or more points of leakage in the defective pigment epithelium (Fig. 1.54). When leakage comes from subretinal neovascularization (Fig. 1.55) or a tumor (Fig. 1.47), it is more rapid and complete. When the pigment epithelium is detached from Bruch’s membrane, fluorescein passes freely and rapidly through Bruch’s membrane from the choriocapillaris into the subpigment epithelial space (Fig. 1.56).

In some cases of central serous chorioretinopathy, there is an associated pigment epithelial detachment, and pooling under each (sensory retinal detachment and the pigment epithelial detachment) is evident. Occasionally, the edge of a pigment epithelial detachment may tear, or rip, and allow fluorescein dye to pass freely into the subretinal space (Fig. 1.57). Drusen may also show late hyperfluorescence similar to that seen with a pigment epithelial detachment (Fig. 1.58). In some cases of pigment epithelial detachment, especially in older patients, subretinal neovascularization is also present. This combination of subretinal neovascularization and pigment epithelial detachment results in an interesting angiogram that can be challenging to interpret (Fig. 1.59).

Fig. 1.57 Late hyperfluorescence under the retina – leakage from the choroid due to a retinal pigment epithelial (RPE) rip. (A) Schematic of a pigment epithelial detachment that has developed a tear along one edge. The barrier function of the pigment epithelium is lost and fluorescein dye can diffuse easily and rapidly in the subretinal space. (B) Color photography of right macula showing a round dark area under the fovea, and light (depigmented area) area extending temporally. In the inferior portion of the macula, some subretinal hemorrhage is seen. (C) Early arteriovenous-phase fluorescein angiogram shows bright hyperfluorescence of the depigmented area temporally, and hypofluorescence under the fovea as well as inferiorly in the area of the subretinal blood. (D) Late-phase fluorescein angiogram shows pooling of fluorescein under the retina where the dye has been able to diffuse freely through Bruch’s membrane in the sensory retinal detachment. Comment: This patient developed a tear of the pigment epithelial detachment. The dark area under the fovea is where the pigment epithelium has rolled up after tearing away from the area temporally. The area temporal to the macula appears light due to absence of the RPE in this area. Since the RPE barrier is absent in this area, the dye diffuses readily and rapidly into overlying sensory retinal detachment, producing late pooling of fluorescein.

Fig. 1.58 Late hyperfluorescent pooling (or staining) of large drusen. (A) Schematic section of retina shows progressively larger detachments of pigment epithelium. Drusen deposit between the pigment epithelium and Bruch’s membrane and lift the pigment epithelium up, forming small or large pigment epithelial detachments, depending on the size of the drusen. (B) Color photograph of right macula shows multiple, pale, round, and variably sized drusen. (C) Arteriovenous phase of fluorescein angiogram shows some early hyperfluorescence of the areas of the drusen. (D) Late phase of fluorescein angiogram shows marked hyperfluorescence of the drusen. The larger drusen take longer for the hyperfluorescence to develop. Comment: The larger the drusen, the more similar they are to pigment epithelial detachments, and therefore the more likely it is that they will show pooling of fluorescein (or staining of the drusen material).

Fig. 1.59 Pigment epithelial detachment with associated (suspicious) subretinal neovascularization. (A) Color photograph of right macula. Note the pigment epithelial detachment temporally and sensory retinal detachment. (B) Arteriovenous-phase fluorescein angiogram shows early hyperfluorescence of the superotemporal pigment epithelial detachment. (C) Late-phase fluorescein angiogram of left macula shows that the fluorescence of the pigment epithelial detachment temporally has increased significantly. Comment: This patient had an irregularly shaped pigment epithelial detachment, which is a sign of possible choroidal neovascularization. The irregular, fuzzy hyperfluorescence is due to likely occult choroidal neovascularization.

In summary, late hyperfluorescence beneath the retina should first be distinguished as pooling of fluorescein into a space or as tissue staining with fluorescein. When pooling is present, one must determine whether a sensory retinal or a pigment epithelial detachment is present. Similarly, if staining is present, one must find out whether the tissue involved is the retinal pigment epithelium and Bruch’s membrane, choroid, or sclera. From this anatomic determination a more specific diagnosis can be determined.

Staining

Staining refers to leakage of fluorescein into tissue or material and is contrasted with pooling of the fluorescein into an anatomic space. Many abnormal subretinal structures and materials can retain fluorescein and demonstrate later hyperfluorescent staining.

Drusen

The most common form of staining occurs with drusen. Most drusen hyperfluoresce early in the angiogram because choroidal fluorescence is transmitted through defects in the pigment epithelium overlying the drusen (Fig. 1.34). Fluorescence from most small drusen diminishes as the dye leaves the choroidal circulation. However, some larger drusen display later hyperfluorescence or staining (Fig. 1.58). The larger the drusen, the more likely they will retain fluorescein and staining will occur. When drusen are large and have smooth edges, the late staining on the angiogram is similar in appearance to that of pooling of fluorescein under a pigment epithelial detachment. In many cases it is difficult, if not impossible, to differentiate large drusen from small pigment epithelial detachments: they have a similar ophthalmoscopic, fluorescein angiographic, and even microscopic appearance.

Scar

Scar tissue retains fluorescein and usually demonstrates well-demarcated hyperfluorescence because little, if any, fluid surrounds the scar. Later in the healing process, when only a few vessels remain, the early angiogram is hypofluorescent because of the paucity of vessels and blockage by the scar tissue. The most commonly seen scar tissue is the disciform scar, which is the endstage of subretinal neovascularization. Scarring is also seen following numerous other insults to the pigment epithelium and choroid, especially inflammation (Fig. 1.60).

Fig. 1.60 Late hyperfluorescence and leakage – staining in geographic helicoid peripapillary choroidopathy (GHPC). (A) Color montage of left disc and macula shows large geographic areas of atrophy of pigment epithelium and choriocapillaris. There is some hyperplasia of the pigment epithelium noted as hyperpigmentation (especially in the macula and papillomacular bundle). Some fibrous scar tissue is present. (B) Arteriovenous-phase fluorescein angiogram shows that the geographic lesions are mostly hypofluorescent; they are caused by loss of pigment epithelium and choriocapillaris. Note that the large choroidal vessels can be seen within these lesions, indicating that the pigment epithelium and choriocapillaris are both gone. There is some hyperfluorescence along the edges of the geographic lesions. The pigment epithelial hyperplasia causes blocked fluorescence. (C) Late fluorescein angiogram of left macula shows hyperfluorescent staining along the edges of the geographic lesion. Comment: This patient had GHPC; inflammation of choroid and pigment epithelium resulted in a loss of the pigment epithelium and choriocapillaris and some of the choroid. The angiogram showed that only large choroidal vessels remained within these lesions. The choriocapillaris was intact, however, in the normal tissue adjacent to the geographic atrophic tissue. The normal choriocapillaris leaked into the atrophic area in a horizontal fashion, causing late hyperfluorescence of areas of scar tissue and some scleral staining.

Sclera

In several situations the sclera is visible ophthalmoscopically and exhibits late hyperfluorescent staining on fluorescein angiography. Scleral staining is best seen when the retinal pigment epithelium is very pale (as in a blonde patient) or when the choriocapillaris is fully intact. When the choriocapillaris is not intact, fluorescein staining of the sclera can occur from the edges of the atrophic area where fluorescein leaks from the intact choriocapillaris inward toward the atrophy (Fig. 1.60).

In conditions such as physiologically light-colored (blonde) fundus or in myopia, the choriocapillaris is usually sufficient to stain the sclera completely. After the choroidal vessels have emptied of fluorescein in the later phases of angiography, the large hypofluorescent choroidal vessels appear as dark lines in silhouette against the stained sclera.

When a loss of choroid and choriocapillaris has occurred, there is a consequent diminution of fluorescein flow in the choroid. When this occurs, the sclera stains with fluorescein only from adjacent normal patent choriocapillaris vasculature. These vessels stain the sclera on the borders of the lesion because the dye tends to diffuse toward the center of the lesion. The entire lesion may not stain if the distance from the edge of the sclera is more than 1 mm. When the choriocapillaris is intact or the lesion is not expansive, the sclera will stain completely.

In summary, late hyperfluorescence beneath the retina should first be distinguished as pooling of fluorescein into a space or as tissue stained with fluorescein. When pooling is present, it must be determined whether a sensory retinal or a pigment epithelium detachment is present. Similarly, if staining is present, it must be determined whether the tissue involved is the retinal pigment epithelium and Bruch’s membrane, choroid, or sclera. From this anatomic differentiation, a more specific diagnosis can be determined.