Abstract
Neovascular age-related macular degeneration (AMD) is characterized by the presence of choroidal neovascularization (CNV) which may result in hemorrhage and edema, hence impairing vision. Fluorescein angiography is used to diagnose neovascular AMD and classify it as classic or occult CNV. Indocyanine-green angiography allows visualization of the CNV lesion and is important in distinguishing typical AMD from polypoidal choroidal vasculopathy. Fundus autofluorescence patterns help to identify abnormal regions of the retina and may be of prognostic value in both nonneovascular and neovascular AMD. Optical coherence tomography identifies the location and extent of the CNV lesion, as well as associated edema and subretinal fluid, and is also useful in monitoring the response to treatment. Newer modalities such as optical coherence tomography angiography will add new dimensions to the diagnosis and monitoring of neovasular AMD.
Keywords
Age-related macular dengeneration, Neovascular, Fluorescein angiography, Fundus autofluorescence, Indocyanine-green angiography, Optical coherence tomography, Optical coherence tomography angiography
Introduction
Age-related macular degeneration (AMD) is a leading cause of blindness worldwide and is the most common cause of blindness in developed countries. The World Health Organization estimates that 8.7% of global blindness is caused by AMD, with 14 million people worldwide rendered either blind or severely visually impaired as a result. Meta-analyses have shown that age-specific prevalence is similar in populations of Asian and Caucasian ancestry, at approximately 6.8%, though early AMD signs are less common among Asians. Factors conferring increased risk include female gender, age, and smoking. An association with cardiovascular risk factors has been identified in various studies, but may not be universal. At present, the prevalence of AMD is increasing further, with patient numbers forecast to increase from 9.1 million in 2010 to 17.8 million in 2050.
The global cost of visual impairment due to AMD is estimated to be US$343 billion, including US$255 billion of direct healthcare cost. The societal costs of AMD are also substantial, in terms of both direct vision-related medical costs (e.g., treatment of AMD and vision-related equipment), direct nonvision-related medical costs (e.g., medications), and direct nonmedical-related costs (e.g., home health care and social services). In addition, severe AMD often causes a decrease in the patient’s quality of life comparable to severe systemic disease.
AMD has traditionally been categorized into neovascular (wet) AMD and nonneovascular (dry) types or presentations. In neovascular AMD, patients manifest with subretinal exudates and hemorrhages, which most commonly originate from choroidal neovascularization (CNV): the proliferation of new capillaries from the choriocapillaris into the subretinal or subretinal pigment epithelium (RPE) space. In contrast, nonneovascular (dry) AMD generally manifests with drusen at the macula and RPE changes, which may not affect vision significantly at an early stage. However, in advanced forms of nonneovascular AMD, such as geographic atrophy of the RPE, patients may also suffer from significant visual impairment. In addition, despite successful pharmacotherapy, many patients with neovascular AMD also eventually go on to develop macular atrophy over time, suggesting that rather than being a distinct type of AMD, wet or neovascular AMD may simply be an interval event in some patients.
The majority of patients with AMD manifest with the nonneovascular form of AMD alone. However, while neovascular AMD accounts for approximately 10–20% of total cases of AMD, it is responsible for 90% of severe visual impairment among AMD patients. Hence, neovascular AMD represents a significant disease.
The current mainstay of treatment for neovascular AMD is antivascular endothelial growth factor (VEGF) agents. These drugs are injected intravitreally, typically at monthly intervals. The efficacy of anti-VEGF agents has been well documented and can result in significant gain of vision from baseline. However, these drugs are also associated with ocular and systemic side effects. Therefore, it is important to accurately diagnose AMD and to monitor its treatment response. Advances in ophthalmic imaging modalities have increased the ability of ophthalmologists to detect the presence of AMD and monitor its progress over time.
Introduction
Age-related macular degeneration (AMD) is a leading cause of blindness worldwide and is the most common cause of blindness in developed countries. The World Health Organization estimates that 8.7% of global blindness is caused by AMD, with 14 million people worldwide rendered either blind or severely visually impaired as a result. Meta-analyses have shown that age-specific prevalence is similar in populations of Asian and Caucasian ancestry, at approximately 6.8%, though early AMD signs are less common among Asians. Factors conferring increased risk include female gender, age, and smoking. An association with cardiovascular risk factors has been identified in various studies, but may not be universal. At present, the prevalence of AMD is increasing further, with patient numbers forecast to increase from 9.1 million in 2010 to 17.8 million in 2050.
The global cost of visual impairment due to AMD is estimated to be US$343 billion, including US$255 billion of direct healthcare cost. The societal costs of AMD are also substantial, in terms of both direct vision-related medical costs (e.g., treatment of AMD and vision-related equipment), direct nonvision-related medical costs (e.g., medications), and direct nonmedical-related costs (e.g., home health care and social services). In addition, severe AMD often causes a decrease in the patient’s quality of life comparable to severe systemic disease.
AMD has traditionally been categorized into neovascular (wet) AMD and nonneovascular (dry) types or presentations. In neovascular AMD, patients manifest with subretinal exudates and hemorrhages, which most commonly originate from choroidal neovascularization (CNV): the proliferation of new capillaries from the choriocapillaris into the subretinal or subretinal pigment epithelium (RPE) space. In contrast, nonneovascular (dry) AMD generally manifests with drusen at the macula and RPE changes, which may not affect vision significantly at an early stage. However, in advanced forms of nonneovascular AMD, such as geographic atrophy of the RPE, patients may also suffer from significant visual impairment. In addition, despite successful pharmacotherapy, many patients with neovascular AMD also eventually go on to develop macular atrophy over time, suggesting that rather than being a distinct type of AMD, wet or neovascular AMD may simply be an interval event in some patients.
The majority of patients with AMD manifest with the nonneovascular form of AMD alone. However, while neovascular AMD accounts for approximately 10–20% of total cases of AMD, it is responsible for 90% of severe visual impairment among AMD patients. Hence, neovascular AMD represents a significant disease.
The current mainstay of treatment for neovascular AMD is antivascular endothelial growth factor (VEGF) agents. These drugs are injected intravitreally, typically at monthly intervals. The efficacy of anti-VEGF agents has been well documented and can result in significant gain of vision from baseline. However, these drugs are also associated with ocular and systemic side effects. Therefore, it is important to accurately diagnose AMD and to monitor its treatment response. Advances in ophthalmic imaging modalities have increased the ability of ophthalmologists to detect the presence of AMD and monitor its progress over time.
Fluorescein Angiography
Fluorescein angiography (FA) is an imaging technique that utilizes the principle of fluorescence, where a substance absorbs light of a specific wavelength (465–490 nm) and reemits it at a longer wavelength (520–530 nm) with lower energy levels. Fluorescein dye is injected intravenously, where around 70–80% of the dye is bound by plasma proteins. Fluorescein does not leak from normal retinal vessels or an intact retinal RPE. However, it does leak from choroidal neovascular membranes such as those which occur in neovascular AMD and can be used to detect the presence and extent of the lesion.
Both FA and indocyanine-green angiography (ICGA) may be performed with modified fundus cameras or confocal scanning laser ophthalmoscopes (CSLOs). With fundus cameras, the entire fundus is illuminated simultaneously by a bright light source, and the reflected light is captured by the charge-coupled device camera. This produces an angiogram in which reflections from various layers of the retina and choroid are superimposed. In contrast, CSLO technology uses a small pinhole to restrict light from a narrow focal plane, thereby enabling an image of a thin layer of the retina without interference from light originating from more superficial or deeper layers.
CNV is a characteristic feature of neovascular AMD. CNV is the result of the growth of new blood vessels from the choroidal circulation which extend into the sub-RPE and/or subretinal space. A classification for CNV was described by Gass based on the anatomical position of the CNV lesion relative to the RPE. Type 1 CNV is located beneath the RPE, in the space between the RPE and Bruch’s membrane. In contrast, Type 2 CNV has penetrated the RPE layer and proliferated in the subretinal (subneurosensory) space. More recently, it has been recognized that neovascularization in the setting of AMD can also originate from the retina (Type 3 CNV) and can grow from an intraretinal location to the subretinal and sub-RPE spaces.
Patterns of CNV on Fluorescein Angiography
The Macular Photocoagulation study established some definitions for the appearance of CNV using FA.
- 1.
Classic CNV . This is defined as an area of uniform hyperfluorescence in the early phase of the angiogram (occurring within the first 40 s or transit phase of the angiogram) ( Fig. 7.1 ). In some cases, particularly in younger patients, the lesion may have a lacy pattern or cartwheel appearance in the early frames and is well demonstrated during dynamic FA. The leakage increases during the mid and late phases of the FA, resulting in an increase in both size and intensity of the hyperfluorescence ( Fig. 7.1C and D ). Although the boundaries are initially well visualized, these become obscured by the intense leakage subsequently.
Figure 7.1
Classic choroidal neovascularization (CNV). (A) Color fundus photograph showing a CNV lesion superior to the fovea. (B) Early phase fluorecein angiogram (FA) demonstrating early leakage. (C) Mid phase FA showing increase in the size and intensity of hyperfluorescence as a result of leakage of fluorescein dye. (D) Late phase fluorescein angiogram. The area of leakage has increased further. (E) Indocyanine green angiogram demonstrating the CNV lesion. (F) Optical coherence tomography with subretinal hyperreflectivity suggestive of Type 2 CNV. There is also retinal pigment epithelium elevation and subretinal fluid above this.
Type 2 CNV (located in the subretinal space) typically appears as a classic CNV on FA. Since the CNV lesion is located above the RPE, the vessels of the CNV lesion may be visualized, and the leakage is usually more intense since there is more room for leakage in the subretinal space.
- 2.
Occult CNV typically corresponds to Type 1 CNV, which is located beneath the RPE ( Fig. 7.2 ). As a result of obscuration by the RPE layer, the individual vessels of the CNV lesion are not well seen, and the pattern of leakage is poorly defined on FA. Two patterns of occult CNV have been described :
- a.
Fibrovascular pigment epithelial detachment (FVPED) . This appears as an area of stippled hyperfluorscence which appears within the first 1–2 min after injection of fluorescein dye. The leakage is usually not as bright or discrete as areas of classic CNV. Using stereoscopic examination, the FVPED has an irregular, rough, or granular appearance. In the late phases, the hyperfluorescence increases as a result of pooling within the pigment epithelial detachment (PED) as well as some spread of dye into the subretinal space, and the boundaries may not be as clear.
- b.
Late leakage of undetermined source (LLUS) . This form of occult CNV is characterized by the appearance of leakage between 2 min and 5 min after fluorescein injection. It often appears as speckled hyperfluorescence, with late pooling of dye. In the early phases of the angiogram, there is no discrete, well demarcated, or discernible area of hyperfluorescence that might be considered the source of leakage. Recently, through FA and optical coherence tomography (OCT) correlation analysis, our group has shown that the main difference between LLUS and FVPED is the thickness of the PED, with LLUS representing a much shallower elevation.
Figure 7.2
Occult choroidal neovascularization (CNV). (A) Color fundus photograph demonstrating the CNV lesion. (B) Early phase fluorescein angiogram (FA) with areas of hyperfluorescence. (C) Mid-phase FA demonstrating leakage superior to the fovea. (D) Late phase FA showing additional leakage superior to the fovea. (E) Indocyanine green angiogram demonstrating the CNV lesion. (F) Optical coherence tomography (OCT) at the fovea, demonstrating pigment epithelium detachment and subretinal fluid. (G) OCT of superior to the fovea, demonstrating pigment epithelium detachment and subretinal fluid.
- a.
Subsequently, the treatment of AMD with photodynamic therapy and verteporfin in photodynamic therapy studies further subdivided CNV lesions into pure classic, predominantly classic, minimally classic or pure occult, based on the amount of each type of CNV relative to other components of the lesion which also included blocked fluorescence, thick blood, and serous PED.
In one series, the frequency of the different types of CNV were 49.6% occult CNV, 12.0% classic CNV, 28.6% retinal angiomatous proliferation (RAP), and 9.8% mixed CNV. Of the mixed CNV lesions, 50.0% were minimally classic, 30.8% predominantly classic, 11.5% occult and RAP, and 7.6% classic and RAP.
Indocyanine-Green Angiography
In contrast to FA, which leaks from the choriocapillaris, indocyanine-green dye is 98% protein-bound and does not leak from the choriocapillaris. This allows better visualization of the choroidal vasculature, as well as abnormal lesions such as CNV lesions and polypoidal choroidal vasculopathy (PCV). In addition, ICGA absorbs light in the near-infrared range (790–805 nm) and has an emission spectrum ranging from 770 to 880 nm, peaking at 835 nm. Because of this longer operating wavelength, ICG is able to fluoresce better through pigment, fluid, lipid, and hemorrhage compared to fluorescein dye, which allows better visualization of the dye through hemorrhage, fluid, or pigments.
In a review by Stanga et al ., the authors recommended ICGA for (1) identification of PCV, (2) in cases of occult CNV, (3) patients with CNV associated with PED, and (4) recurrent CNV membranes. In these conditions, ICGA helped to confirm the diagnosis and identify treatable lesions or feeder vessels.
Among patients with neovascular AMD, ICGA is useful to detect CNV lesions. ICG angiography is often employed in conjunction with FA and aids to confirm the FA findings of CNV in patients with well defined CNV. It is especially useful in cases of occult CNV, where FA is unable to visualize the neovascular net or in patients where the neovascularization may be blocked by the presence of hard exudates or hemorrhages.
In a study of 51 patients with acute spontaneous submacular hemorrhage, Kim et al . reported that the cause of the hemorrhage was diagnosed in 84.3% of eyes based on ICGA findings, with 93% of the initial diagnoses being correct. In three patients, however, an initial diagnosis of AMD was revised to PCV on follow-up ICGA. In that series, the most common causes of submacular hemorrhage were neovascular AMD (52.9%) and PCV (37.3%).
Morphological Appearance of CNV on ICGA
CSLO, which separates the illuminating beam and the imaging beam in the eye, can be used for high-speed ICG angiography. This allows for the visualization of CNV lesions and feeder vessels through dynamic ICG angiography. In studies on eyes with occult CNV identified using FA, digital ICG videoangiography identified three types of CNV lesions: (1) focal spots, (2) plaques (well defined or poorly defined), and (3) combination lesions (containing both focal spots and plaques). The combination lesions were further subdivided into marginal spots (where the focal spots were situated at the edge of plaques of neovascularization), overlying spots (hot spots located over plaques of neovascularization), and remote spots (a focal spot which was remote from a plaque of neovascularization). In one study of 244 consecutive patients with occult CNV, and an associated serous PED, ICGA identified focal CNV in 38% of eyes, whereas the remaining 62% were defined as plaques. Thus, ICGA identified patients who were potentially eligible for focal laser photocoagulation.
ICGA-Guided Treatment of CNV Lesions
Before the widespread use of anti-VEGF injections, focal laser photocoagulation was useful in the treatment of some cases of neovascular AMD. In many eyes, however, the lesion was deemed to be too extensive based on the findings on FA. ICGA served to more clearly delineate the CNV lesion and allowed some cases which were previously deemed to be untreatable to be successfully treated. In a series of 23 eyes with untreated CNV secondary to neovascular AMD, ICGA identified focal spots at the edge of a plaque. Following ICGA-guided laser photocoagulation, anatomic success with resolution of the exudative findings was achieved in 79% of eyes at 6 months and 68% of eyes at 12 months.
In a retrospective review of 252 consecutive patients with neovascular AMD, it was found that early examination of eyes using ICGA allowed identification of lesions which were amenable to treatment with PDT with verteporfin. Among eyes examined within 15 days, 49% had focal spots compared to 32% of eyes examined between 16 days and 30 days.
ICGA was also used to identify feeder vessels supplying the CNV lesions. In a study of 170 consecutive patients with subfoveal CNV, 37 patients manifested with feeder vessels which were treated with laser photocoagulation with resolution of exudative changes in 70% of cases.
Predicting Response to Treatment
More recently, intravitreal injections of anti-VEGF drugs have become the mainstay of treatment of neovascular AMD. ICGA has been shown to be useful in predicting which patients are more likely to experience recurrent exudation and/or subretinal hemorrhage after treatment extension. In a study of patients with neovascular AMD being treated with bevacizumab using a treat-and-extend regimen, patients whose treatment intervals could not be extended manifested with an increase in CNV area of 33% or more based on ICGA, which was significantly greater than the group whose treatment intervals were successfully extended.
ICGA in Variants of Neovascular AMD
A variant of neovascular AMD is PCV, which is characterized by an abnormal vascular network, often referred to as a branching vascular network, and terminal dilatations which form the polyps ( Fig. 7.3 ). First described as peripapillary lesions among African-American females, PCV is now known to occur more commonly among certain populations, such as Asians, where it may account for up to 55% of patients presenting with neovascular AMD.
Since the features of PCV on FA appear similar to neovascular AMD, ICGA is essential for the accurate diagnosis of PCV. Although different diagnostic criteria are used by various investigators, the standardized criteria used in the EVEREST study consisted of early, focal hyperfluorescence on ICGA, appearing within the first 5 min, together with at least one of the following criteria:
- 1.
Nodular appearance of the polyps on stereoscopic examination,
- 2.
Presence of a hypofluorescent halo around the nodule,
- 3.
Presence of abnormal vascular channels supplying the polyps,
- 4.
Pulsation of the polyp,
- 5.
Orange-red subretinal nodule(s) which correspond to the location of the hyperfluorescence on ICGA, and
- 6.
Massive submacular hemorrhage (defined as hemorrhage with an area of four disc areas or larger).
The use of CSLO ICGA is useful because the polyps and the branching vascular networks are shown more clearly compared to conventional flash ICGA ( Fig. 7.3E ). In addition, the presence of pulsation of the polyp can only be detected during the dynamic phase of the angiogram.
In the EVEREST study, the frequency of the various diagnostic features were nodular appearance—91.8%, abnormal vascular channels—88.5%, hypofluorescent halo—68.9%, pulsation of the polyp—6.6%, orange subretinal nodules—55.7%, and massive submacular hemorrhage—13.1%.
Some papers have described subtypes of PCV, based on differences in their appearance on ICGA. In one study, three subtypes of PCV were described based on the type of abnormal vascular channel supplying the polyps (a series of interconnecting channels or a branching vascular network) and the presence of significant leakage on FA. Type A PCV had interconnecting channels only, Type B had a branching vascular network but no significant leakage on FA, while Type C had a branching vascular network with significant leakage on FA. Over 5 years, the visual outcomes were best for those with Type A, intermediate for those with Type B, and worst for those with Type C. This classification may be useful in prognosticating patients and identifying those who may require closer follow-up and more aggressive treatment.
Fundus Autofluorescence
Autoflourescence is a phenomenon where tissues demonstrate fluorescence without the use of a dye (such as fluorescein). Fundus autofluorescence (FAF) is an imaging modality for the metabolic mapping of naturally or pathologically occurring ocular fluorophores such as lipofuscin. Lipofuscin is formed from the phagocytosis of damaged photoreceptor outer segments by the RPE cells. Over time, as RPE phagocytose more outer segments, lipofuscin accumulates in the cell. Other causes for lipofuscin accumulation include oxidative stress and disease states. Areas of hyperfluorescence indicate excess lipofuscin accumulation, while areas of hypofluorescence may indicate RPE cell death. Hence, lipofuscin accumulation may be an early marker of retinal degeneration. FAF is useful in AMD as it may show phenotypes that are not evident on color photography or other imaging modalities.
FAF can be imaged using green (532 nm) and blue (488 nm) light. A standard fundus camera such as the Topcon TRC-50DX uses flash photography with modified green light filters (Spaide AF filters), while the Heidleberg Spectralis uses CSLO with blue light filters to obtain FAF images.
In CNV, a typical FAF pattern consists of various patterns of hypoautofluorescence corresponding to hemorrhages, exudates, and atrophy. There may also be hyperautofluorescence in areas with RPE proliferation and lipofuscin accumulation. In late stages of AMD, where there is a disciform scar, FAF shows a hypoautofluorescent area corresponding to the scar. Some scars may also have some increased FAF signal at the junctional zone.
Some studies have postulated that the accumulation of lipofuscin, and thereby the presence of autofluorescence, especially at the junctional area of geographic atrophy, precedes cell death. Holz et al. reported that in patients with geographic atrophy, areas of increased autofluorescence heralded new atrophic areas with enlargement of existing atrophic areas. In particular, areas of increased autofluorescence preceded the development and enlargement of outer retinal atrophy. This may be due to a direct pathogenetic effect of lipofuscin causing RPE dysfunction, or that excessive accumulation of lipofuscin already signifies RPE cell dysfunction.
Classification of Abnormal FAF Patterns in AMD
The FAF in AMD (FAM) Study proposed a classification for FAF patterns in AMD:
- 1.
Focal increased FAF at the margin of the atrophic patch,
- 2.
Continuous stippled band of increased FAF surrounding the entire atrophic area,
- 3.
Patchy increased FAF outside the area of GA. FAF intensity of these patches tended to be of lesser degree compared with the focal pattern, and
- 4.
Diffuse increase in FAF. This was further subdivided into
- a.
Reticular —Various lines of increased FAF with a preferred radial orientation,
- b.
Branching —Fine branching pattern of an increased FAF signal,
- c.
Fine granular —Larger area of increased FAF with a granular-like appearance surrounding the GA, and
- d.
Fine granular with peripheral punctate spots —Diffuse FAF change surrounding the atrophic area with elongated small lesions with increased FAF signal.
- a.
Correlation of FAF Abnormalities With Clinical Features
The patterns of autofluorescence differ from non-AMD-related geographic atrophy. In a study comparing 57 eyes of patients with GA secondary to AMD and 43 eyes with GA secondary to other etiologies (such as juvenile macular dystrophies), increased autofluorescence outside the area of GA was observed in 82.5% of eyes with AMD compared to 9.3% of the non-AMD group ( p < 0.001).
In eyes with bilateral AMD, the patterns of autofluorescence changes have been reported to be identical in 89.5% of patients. This was also reported in studies of peripheral autofluorescence changes in eyes with AMD, suggesting the possibility of a genetic influence on the abnormal FAF patterns.
Some authors have reported clear relationships between the patterns of abnormal FAF and the clinical findings in eyes with AMD. For example, a study reported that in early AMD, focally increased autofluorescence colocalized with large, soft drusen, and hyperpigmentation. In advanced atrophic AMD, however, this pattern was found adjacent to drusen and regions of GA. In addition, reticular hypoautofluorescence was reported to be a highly sensitive marker for reticular pseudodrusen and was strongly associated with CNV. Another longitudinal study of 221 eyes reported that areas of GA appeared as mixed autofluorescence, whereas regions of drusen-associated atrophy appeared as hypoautofluorescence.
In contrast, other studies suggested that regions of abnormal FAF signals may not always correspond to fundoscopically visible changes in the retina. In a prospective randomized controlled trial of prophylactic laser therapy in patients with AMD, Lois et al . reported that there was no obvious correspondence between the FAF pattern and the distribution of drusen, and suggested that FAF abnormalities and drusen may represent independent measures of aging in the posterior pole.
Rate of Progression of Different FAF Patterns and Prognostic Significance
Studies have shown that the type of abnormal FAF in AMD may have prognostic significance and manifest with differences in rates of progression. In a study of 195 eyes, Holz et al . found that progression rates in eyes with banded and diffuse FAF patterns were significantly higher compared to eyes without FAF abnormalities or with focal FAF patterns. Among those with the diffuse pattern, eyes with a diffuse trickling pattern had higher rates of progression of GA compared to the others. Studies have reported that the baseline area of FAF abnormality was a predictor of the rate of increase in GA.
Abnormal FAF patterns may also have systemic correlations. Patients with diffuse-trickling FAF were younger, more commonly male, and had a higher rate of systemic cardiovascular disorders in the younger age group (including myocardial infarction, hypertensive crisis, and angina) compared to those with nondiffuse-trickling patterns. The authors suggested that vascular insufficiency at the level of the choroid may play a pathogenetic role in patients with this particular pattern of abnormal FAF.
FAF in Eyes With Neovascular AMD
Knowledge of the patterns of FAF in eyes with geographic atrophy are relevant in patients with neovascular AMD. Even though geographic atrophy was previously thought to be a feature of nonneovascular AMD, it is now evident that atrophy also occurs in eyes with neovascular AMD.
In the SEVEN-UP study, which reviewed patients from the ANCHOR, MARINA, and HORIZON trials 7 years after enrollment, 57 of 58 eyes (98%) had atrophy, of which 90% involved the fovea. A decrease in best corrected visual acuity at the SEVEN-UP visit was associated with both increased area of macular atrophy ( p =0.0001) and the presence of subfoveal macular atrophy ( p < 0.01). For every 1-mm increase in macular atrophy, there was a 2.3 letter decrease in BCVA ( p < 0.0001). In multivariate regression analyses, macular atrophy at year 7 was the only anatomic factor associated significantly with final vision ( p < 0.001). For each 1 mm 2 of macular atrophy, there was an associated 1.6 letter deficit in ETDRS vision score at the SEVEN-UP visit.
A study of the mean gray scale on FAF images found a significant increase in values following treatment for both types 1 and 2 CNV. It was postulated that decreases in subretinal and intraretinal fluid (IRF) may be a cause. In addition, it is believed that the increase in posttreatment FAF may indicate the presence of preserved RPE and photoreceptor function, which implies a favorable long-term visual prognosis.
Several studies have described several patterns of FAF in eyes with neovascular AMD.
- •
Normal or slightly decreased FAF . These are seen typically in early lesions, where the RPE function has been preserved.
- •
Decreased FAF at the center of the lesion, with increased FAF signal at the lesion edge ( Fig. 7.4 ). This has been described to occur in patients with classic CNV and is believed to be caused by the damage to the RPE cells caused by the proliferation of the CNV lesion between the neurosensory retina and the RPE. The location of the CNV lesion above the level of the RPE would also contribute to the blockage of the FAF signal.
Figure 7.4
Neovascular age-related macular degeneration. (A) Color fundus photography demonstrating the CNV lesion and subretinal hemorrhage surrounding the CNV lesion. (B) Fundus autofluorescence with hypoautofluorescence at the fovea, with surrounding hyperautofluorescence. (C) Early phase fluorescein angiogram (FA) demonstrating leakage. There is an area of blocked fluorescence due to the presence of subretinal hemorrhage. (D) Late phase FA demonstrating leakage from the CNV lesion, with both classic and occult CNV seen. (E) Indocyanine green angiogram of the CNV lesion. (F) Optical coherence tomography demonstrating hyperreflectivity above the retinal pigment epithelium. There is subretinal fluid and intraretinal cysts.
- •
Spotty and heterogeneous FAF patterns. This feature has been described in patients in occult CNV and is believed to be due to the segmental and patch loss of RPE caused by the slow growth of the CNV lesion beneath the RPE.
- •
Hyperautofluorescent ring . This has been described in some eyes and occurs at the margins of the CNV complex and may be related to the extent of the subretinal fluid (SRF). It has also been found that eyes with a hyperautofluorescent ring were more likely to have loss or disruption of the ellipsoid zone compared to eyes without this feature.
Ultra-Widefield Autofluorescence in AMD
In recent years, the introduction of ultra-widefield imaging devices has allowed imaging of the peripheral retina, covering up to 200 degrees in a single image. Studies have reported patterns of peripheral FAF abnormalities in eyes with AMD. Among the patterns, granular hyperautofluorescence, mottled, and nummular hypoautofluorescence have been described. Peripheral FAF abnormalities are more common in eyes with neovascular AMD compared to nonneovascular AMD or normal eyes (86% vs 72.8% vs 18.4%). Significant risk factors included the presence of neovascular AMD, older age, and female sex.
The types of peripheral FAF abnormalities correlate with clinical features. For example, granular FAF was associated with peripheral drusen, CNV, or GA ; mottled FAF with RPE depigmentation and nummular decreased autofluorescence with atrophic patches.
Fluorescein Angiography
Fluorescein angiography (FA) is an imaging technique that utilizes the principle of fluorescence, where a substance absorbs light of a specific wavelength (465–490 nm) and reemits it at a longer wavelength (520–530 nm) with lower energy levels. Fluorescein dye is injected intravenously, where around 70–80% of the dye is bound by plasma proteins. Fluorescein does not leak from normal retinal vessels or an intact retinal RPE. However, it does leak from choroidal neovascular membranes such as those which occur in neovascular AMD and can be used to detect the presence and extent of the lesion.
Both FA and indocyanine-green angiography (ICGA) may be performed with modified fundus cameras or confocal scanning laser ophthalmoscopes (CSLOs). With fundus cameras, the entire fundus is illuminated simultaneously by a bright light source, and the reflected light is captured by the charge-coupled device camera. This produces an angiogram in which reflections from various layers of the retina and choroid are superimposed. In contrast, CSLO technology uses a small pinhole to restrict light from a narrow focal plane, thereby enabling an image of a thin layer of the retina without interference from light originating from more superficial or deeper layers.
CNV is a characteristic feature of neovascular AMD. CNV is the result of the growth of new blood vessels from the choroidal circulation which extend into the sub-RPE and/or subretinal space. A classification for CNV was described by Gass based on the anatomical position of the CNV lesion relative to the RPE. Type 1 CNV is located beneath the RPE, in the space between the RPE and Bruch’s membrane. In contrast, Type 2 CNV has penetrated the RPE layer and proliferated in the subretinal (subneurosensory) space. More recently, it has been recognized that neovascularization in the setting of AMD can also originate from the retina (Type 3 CNV) and can grow from an intraretinal location to the subretinal and sub-RPE spaces.
Patterns of CNV on Fluorescein Angiography
The Macular Photocoagulation study established some definitions for the appearance of CNV using FA.
- 1.
Classic CNV . This is defined as an area of uniform hyperfluorescence in the early phase of the angiogram (occurring within the first 40 s or transit phase of the angiogram) ( Fig. 7.1 ). In some cases, particularly in younger patients, the lesion may have a lacy pattern or cartwheel appearance in the early frames and is well demonstrated during dynamic FA. The leakage increases during the mid and late phases of the FA, resulting in an increase in both size and intensity of the hyperfluorescence ( Fig. 7.1C and D ). Although the boundaries are initially well visualized, these become obscured by the intense leakage subsequently.
Figure 7.1
Classic choroidal neovascularization (CNV). (A) Color fundus photograph showing a CNV lesion superior to the fovea. (B) Early phase fluorecein angiogram (FA) demonstrating early leakage. (C) Mid phase FA showing increase in the size and intensity of hyperfluorescence as a result of leakage of fluorescein dye. (D) Late phase fluorescein angiogram. The area of leakage has increased further. (E) Indocyanine green angiogram demonstrating the CNV lesion. (F) Optical coherence tomography with subretinal hyperreflectivity suggestive of Type 2 CNV. There is also retinal pigment epithelium elevation and subretinal fluid above this.
Type 2 CNV (located in the subretinal space) typically appears as a classic CNV on FA. Since the CNV lesion is located above the RPE, the vessels of the CNV lesion may be visualized, and the leakage is usually more intense since there is more room for leakage in the subretinal space.
- 2.
Occult CNV typically corresponds to Type 1 CNV, which is located beneath the RPE ( Fig. 7.2 ). As a result of obscuration by the RPE layer, the individual vessels of the CNV lesion are not well seen, and the pattern of leakage is poorly defined on FA. Two patterns of occult CNV have been described :
- a.
Fibrovascular pigment epithelial detachment (FVPED) . This appears as an area of stippled hyperfluorscence which appears within the first 1–2 min after injection of fluorescein dye. The leakage is usually not as bright or discrete as areas of classic CNV. Using stereoscopic examination, the FVPED has an irregular, rough, or granular appearance. In the late phases, the hyperfluorescence increases as a result of pooling within the pigment epithelial detachment (PED) as well as some spread of dye into the subretinal space, and the boundaries may not be as clear.
- b.
Late leakage of undetermined source (LLUS) . This form of occult CNV is characterized by the appearance of leakage between 2 min and 5 min after fluorescein injection. It often appears as speckled hyperfluorescence, with late pooling of dye. In the early phases of the angiogram, there is no discrete, well demarcated, or discernible area of hyperfluorescence that might be considered the source of leakage. Recently, through FA and optical coherence tomography (OCT) correlation analysis, our group has shown that the main difference between LLUS and FVPED is the thickness of the PED, with LLUS representing a much shallower elevation.
- a.
