Figure 7.1 Representative images of polypoidal choroidal vasculopathy (PCV). Pigment epithelial detachment (A) and subretinal hemorrhage (B) are common manifestations. C. Subretinal lipid is often observed. D. In some patients, progressive subretinal fibrosis can be observed. (Reprinted from Imamura Y, Engelbert M, Iida T, et al. Polypoidal choroidal vasculopathy: a review. Surv Ophthalmol. 2010;55(6):501–515. doi: 10.1016/j.survophthal.2010.03.004. Epub 2010 Sep 20, with permission.)
PCV has been classified clinically (30) as (i) quiescent, polyps in the absence of subretinal or intraretinal fluid or hemorrhage; (ii) exudative, exudation without hemorrhage, which includes variously neurosensory retinal thickening, neurosensory detachment, PED, and subretinal lipid exudation; and (iii) hemorrhagic, any subretinal or sub-RPE hemorrhage with or without other exudative characteristics. This clinical classification described the primary features of the disease, but the evidence regarding its correlation with prognosis or treatment selection is limited.
The typical manifestations in a patient who is symptomatic for less than 3 months are extensive subretinal exudation and bleeding with minimal cystic change in the retina and a surprisingly good visual acuity (VA). This discrepancy between the severity of the serosanguineous detachments and good visual acuity is best explained by the minimal intraretinal changes. For patients with symptoms longer than 3 months, there are considerable lipid depositions from protein leakage from active aneurysmal elements in the polypoidal vascular abnormalities (6) (Fig. 7.1C).
The morphology of PCV can be distributed into three groups (i) single polyp (Fig. 7.2A); (ii) cluster: more than two polyps in a group (Fig. 7.2B); and (iii) string: three or more polyps in a line (Fig. 7.2C).
Figure 7.2 Midphase ICGA revealing: A. Single polyp. B. Cluster of polyps. C. String of polyps. (Reprinted from Cackett P, Wong D, Yeo I. A classification system for polypoidal choroidal vasculopathy. Retina. 2009;29(2):187–191. doi: 10.1097/IAE.0b013e318188c839, with permission.)
Cackett et al. (31) found cluster formation in 66.7% of the cases, single in 27.5%, and string in 5.8% in a retrospective study of 123 patients.
Uyama et al. (32) found that PCV lesions resembling a cluster of grapes had marked bleeding and leakage and high risk of severe visual loss.
It therefore seems that clusters are associated with an increased frequency of massive hemorrhage resulting in poorer visual outcomes.
NATURAL HISTORY OF THE DISEASE
The natural history of PCV depends on factors including location (peripapillary vs. macular), size of the lesion, and associated bleeding and exudation—which may resolve or progress, sometimes to extensive subretinal fibrosis. The evolution of this condition is probably heavily influenced by the racial background and individual genetic makeup. In Japanese patients, PCV is a chronic disease with a variable course. The risk factors seen in CNV secondary to AMD, including soft confluent drusen and focal hyperpigmentation, are not notable findings in patients with PCV.
Despite having recurrent leakage and bleeding, not all patients develop disciform scars, and others may retain useful vision for years. Some patients, however, experience severe visual loss from a combination of extensive bleeding, exudation, and macular damage.
In a study of the natural history of patients with PCV, Uyama et al. (32) reported that half of the study eyes had hemorrhagic episodes, recurrent leakage, or severe RPE atrophy after a long follow-up period (24–54 months) (33). Following the spontaneous resolution of the acute serosanguineous complications, there may be signs of subretinal fibrosis, pigment epithelial hyperplasia, and atrophic degeneration (34) (Fig. 7.1D). The incidence of sub-RPE hemorrhage or subretinal hemorrhage described in patients with PCV is high (30%–64%) (35,36).
On the other hand, Uyama et al. (32) described stabilization of visual acuity and regression of neovascular lesion with good visual prognosis in 50% of eyes with polypoidal CNV in the posterior pole that were followed for at least 2 years.
PCV has been described in association with other macular abnormalities, such as central serous chorioretinopathy (37,38), typical neovascular (type 1 or 2) AMD (39), sickle cell retinopathy (40), melanocytoma of the optic nerve (41), circumscribed choroidal hemangioma (42), tilted disk syndrome (43), pathological myopia (43), choroidal osteoma (44), and angioid streaks in pseudoxanthoma elasticum (45).
Imaging such as fluorescein angiography (FA), ICGA, and high-resolution optical coherence tomography (OCT) is important for the diagnosis of PCV.
The EVEREST study was the first to include a set of well-defined criteria for PCV diagnosis.
This study defined PCV as the presence of early subretinal focal ICGA hyperfluorescence (appearing within the first 6 minutes after injection of indocyanine green) and in addition, at least one of the following angiographic or clinical criteria: (i) association with a branching vascular network (BVN) seen in ICGA; (ii) presence of pulsatile polyp seen in dynamic ICGA; (iii) nodular appearance when viewed stereoscopically; (iv) presence of hypofluorescent halo (in first 6 minutes); (v) orange subretinal nodules in stereoscopic color fundus photograph (polyp corresponding to ICGA lesions); or (vi) association with massive submacular hemorrhage (defined as size of hemorrhage of at least 4 disk areas) (33).
Fundus autofluorescence (FAF) photography is used to visualize lipofuscin, which accumulates in RPE and provides information about RPE metabolism and function.
Yamagishi et al. (46), studied the incidence and distribution of hypoautofluorescence comparing Japanese patients with PCV and typical neovascular AMD.
In affected eyes with PCV, the sites of the neovascular lesions showed two distinct FAF patterns: (i) confluent hypoautofluorescence at the polypoidal lesions and (ii) granular hypoautofluorescence at the branching choroidal vascular networks.
Confluent hypoautofluorescence was defined as a manifestation of a homogeneous lack of autofluorescence that was well demarcated and clearly distinguishable from the other adjacent lesions. Granular hypoautofluorescence was defined as a heterogeneous mixed finding of the hypoautofluorescence lesion at the various levels.
The confluent hypoautofluorescence pattern was seen in 80.4% eyes with PCV, while it was seen in no eyes with AMD, suggesting that it is a pattern exclusively of PCV. The granular hypoautofluorescence pattern was seen in both PCV (98.9%) and AMD (87.1%) (46).
Overall the incidence of hypoautofluorescence was higher in the PCV group.
The difficulty in precisely delineating structures under the RPE has limited the usefulness of FA in studying and diagnosing the disease.
However, typical findings show mottled hyperfluorescence in the early stage of the angiogram with late staining of the polyps (47), associated with serous or serosanguineous pigment epithelial detachment (PED). Serous PED in the early arteriovenous phase of the angiogram demonstrates progressive, uniform hyperfluorescence with late, intense pooling of fluorescein, while hemorrhagic PED blocks fluorescence and the normal choroid. Alternatively, PCV can be obscured by greater fluorescence from staining serous PED or blocked by serosanguineous PED, making it impossible to determine the exact boundary of the lesion.
This serosanguineous detachment may or may not be vascularized (48).
In many clinical trials, FA is performed to determine CNV activity.
Figure 7.3 ICGA corresponding with large BVN ending in string polypoidal lesion. Identification of a BVN is not an absolute requirement for a diagnosis of PCV because it is not always clearly visualized by ICGA. (Reprinted from Koh AH, et al. Polypoidal choroidal vasculopathy: evidence-based guidelines for clinical diagnosis and treatment. Retina. 2013;33(4):686–716. doi: 10.1097/IAE.0b013e3182852446, with permission.)
Indocyanine Green Angiography
ICGA is essential for a definitive diagnosis of PCV (1,6,50). The longer wavelengths used in this imaging system penetrate more deeply. The ICG molecule is stimulated by the absorption of infrared light in the range from 790 to 805 nm. The RPE and the choroid absorbs up to 75% of the blue-green light used for FA, but only up to 38% of the near infrared light used for ICG. The higher transmittance of light above 800 nm and the strong intravascular retention of the ICG molecule allow better resolution of the choroidal vasculature.
PCV is defined as the presence of single or multiple focal areas of hyperfluorescence arising from the choroidal circulation within the first 6 minutes after injection of indocyanine green, with or without an associated BVN.
The recommended time window in which the polyps of PCV appear after injection of indocyanine green is given as 6 minutes, as used by the EVEREST trial.
In the early phases of the ICGA, dye is visible within the vascular network of the PCV lesion, before it is visible in the retinal vessels. Several seconds after the network vessels are seen to fill with dye, small hyperfluorescent “polyps” became visible. These vascular polypoidal structures usually occur at the termini of vessels at the edge of the vascular network. The area within and immediately surrounding the network remains hypofluorescent compared to the uninvolved choroid. The polypoidal structures can show leakage of ICG dye, with the dye collecting initially in the core of the polypoidal structure. In the late phase, the core of the polypoidal structure becomes relatively hypofluorescent because of the washout of the dye, producing a ring-like staining of the polyps. This finding may help in distinguishing between the polyps that are actively leaking, which will become visible as a ring of hyperfluorescence, and those that are not, which will appear as hypofluorescent rings (51).
Lesion size seems to influence the fluorescence pattern. Lesions smaller than half a disk diameter appear to have intense uniform fluorescence, whereas internal details seem to be visible in larger polypoidal lesions, suggesting the presence of an internal architecture.
The appearance of vessels in PCV often depend on their location in the fundus. In juxtapapillary lesions, the vascular channels may follow a radial, arching pattern and may be interconnected with smaller spanning branches more evident and numerous at the edges of the lesion. When PCV is limited to the macula, a vascular network often follows an oval distribution. With macular and juxtapapillary involvement, vessels in the network usually course in an irregular latticework and do not follow the lobular pattern of choroidal vasculature.
Because the choriocapillaris cannot be clearly visualized with ICGA, the precise location of the polypoidal structures in relation to the choriocapillaris cannot be stated with certainty.
Optical Coherence Tomography
Both time- and spectral-domain OCT (SD-OCT) have confirmed that pathologic vascular lesions of PCV are located beneath the RPE. They appear on OCT scans as abnormalities in the contour of the highly reflective line representing the RPE (53,54). The polypoidal lesions appear as sharp protrusions of the RPE with moderate reflectivity beneath the RPE line (27,55).
Using SD-OCT, moderate reflectivity has been noted in the space between undulating RPE and straight Bruch’s membrane lines. These areas seem to correspond with BVN.
In conventional OCT scans, BVN were represented by one or two highly reflective blurred lines, where the outer reflective layer is believed to represent the inner boundary of the Bruch’s membrane–choriocapillaris complex.
In some of these eyes, cross-sectional images through the PCV lesions resembled the appearance of “pearls on a string” as described by Freund et al. (56), however, polyp structures are not always visible by OCT (6) (Fig. 7.4A and B).
FIGURE 7.4 A,B.. SD-OCT B-scan shows multiple polyps structures adherent to the undersurface of a RPE detachment (arrows) creating “pearls on a string.” C. SD-OCT B-scan shows PCV lesions (arrows) located between the RPE and Bruch’s membrane creating a “double-layer sign.” D. SD-OCT B-scan shows multiple PCV lesions adherent to the undersurface of a RPE detachment. The arrow shows part of the Bruch’s membrane that remains attached to the undersurface of the PCV lesions, creating a “triple-layer sign.” (Reprinted from Khan S, Engelbert M, Imamura Y, et al. Polypoidal choroidal vasculopathy: simultaneous indocyanine green angiography and eye-tracked spectral domain optical coherence tomography findings. Retina. 2012;32(6):1057–1068. doi: 10.1097/IAE.0b013e31823beb14, with permission.)
The PCV lesions can be found both at the margin and within the associated type 1 neovascular lesions (6). The PCV lesions are localized to just below the hyperreflective RPE band near the anterior surface of the associated type 1 neovascular tissue (56).
They hypothesized that the “double-layer” sign occurs when the RPE and its basement membrane (“inner reflective layer”) become separated from the remainder of Bruch’s membrane and the inner choroid (“outer reflective layer”) by fluid.
Abe et al. (58) demonstrated that the OCT “double-layer” sign is most prominent in areas where ICGA shows “abnormal blood vessel networks” in association with larger PCV lesions.
Khan et al. (56) described a “triple-layer sign” when part of the Bruch’s membrane had detached from the underlying choroid (Fig. 7.4D); this separation creates a second hyporeflective space beneath the elevated remainder of Bruch’s membrane, suggesting that the polyps and BVN are more adherent to the basal surface of the RPE than to the underlying choroid, which supports the hypothesis that PCV represents a form of neovascular tissue rather than alterations of the native choroidal vasculature.
PCV may be considered as active if there is clinical OCT or FA evidence of any one of the following that is attributable to PCV: Vision loss ≥5 letters (Early Treatment Diabetic Retinopathy Study) or equivalent, subretinal fluid or intraretinal fluid, PED, subretinal or sub-RPE hemorrhage, and fluorescein leakage. Active PCV can be symptomatic or asymptomatic: treatment should be initiated for active and symptomatic PCV and can be considered for active, asymptomatic PCV (30).
Direct thermal laser photocoagulation still has a role in the treatment of extrafoveal polyps. The ICGA-guided thermal laser photocoagulation may be considered for extrafoveal polyps. However, direct thermal laser photocoagulation is not recommended for the initial therapy of active juxtafoveal or subfoveal PCV given the destructive nature of this treatment modality (30).
In 2003, Yuzawa et al. (61) noted that total lesion ablation provided better outcomes compared with laser ablation of the vascular polyps alone. They evaluated the efficacy of photocoagulation in 47 eyes of PCV. Of the 10 eyes undergoing photocoagulation of whole lesions, 9 showed absorption of exudate and/or blood and maintained or improved visual acuity. However, of the 37 eyes undergoing laser photocoagulation of only the polypoidal component, 20 (54%) showed decreased visual acuity because of recurrent or persistent exudation or due to foveal atrophy.
They concluded that photocoagulation is recommended only if it is feasible to treat the entire polypoidal lesion.
Photodynamic therapy (PDT) is currently the most successful treatment reported.
In the EVEREST trial, 61 Asian patients with symptomatic PCV were treated with verteporfin PDT monotherapy, 0.5 mg ranibizumab monotherapy, or a combination of these treatments. The primary endpoint was the proportion of patients with ICGA-assessed complete regression of polyps at 6th month. The secondary endpoints included mean change in best-corrected visual acuity (BCVA) at 6th month and safety.
PDT was administered according to the treatment protocol of the Age-Related Macular Degeneration with Photodynamic Therapy study (62); patients were infused with verteporfin (6 mg/m2), and 15 minutes after the start of infusion, PDT at standard fluence (light dose, 50 J/cm2; dose rate, 600 mW/cm2; wavelength, 689 nm) was applied to the study eye for 83 seconds. The laser spot size was derived by adding 1,000 mm to the greatest linear dimension. Thus, both BVN and polyps were included in the verteporfin PDT treatment area.
They showed that at 6 months, verteporfin combined with ranibizumab or alone was superior to ranibizumab monotherapy achieving complete polyp regression (77.8%, 71.4%, and 28.6%, respectively). Also, they showed that mean change (±standard deviation) in best-corrected visual acuity was 10.9 (±10.9) for the verteporfin PDT + ranibizumab group, 7.5 (±10.6) for verteporfin PDT alone, and 9.2 (±12.4) for ranibizumab monotherapy. They concluded that verteporfin PDT combined with ranibizumab 0.5 mg or alone were superior to ranibizumab monotherapy in achieving complete regression of polyps. Because the study was not designed to demonstrate a difference between verteporfin PDT plus ranibizumab combination therapy and verteporfin PDT monotherapy, the evidence supports the use of either therapy (30).
Yannuzzi et al. (48) evaluated the efficacy of “selective PDT” with ICGA-guided PDT with verteporfin for PCV.
They used ICGA because it provides improved imaging of the true extent of the PCV lesion without obscuration by blood or leakage at the site of serous pigment epithelial elevations. This method avoided treating the nonvascular components of the lesion mandated by standard FA-based techniques. In addition, the smaller PDT spot size minimized the possibility of choroidal ischemic changes and, therefore, stimulation of vascular endothelial growth factor (VEGF) (63–66).
They applied standard PDT with verteporfin with a spot size of 200 μm around the active component adding 400 μm to the greatest linear dimension. They avoided the fovea if it was not affected.
PDT with verteporfin was applied using the same parameters as for the treatment of neovascular AMD in the Treatment of Age-Related Macular Degeneration with Photodynamic Therapy Study (62). Retreatment was performed under ICGA guidance when leakage was present in FA.
Their outcomes were improvement of vision ≥3 ETDRS lines in 15 eyes (50%), decrease of vision with a loss of ≥3 ETDRS lines in 6 eyes (20%), and stable vision in 9 eyes (30%).
Massive subretinal hemorrhage has been treated by pars plana vitrectomy, with poor visual outcomes (53).
Retinal function as assessed by multifocal electroretinograms can be disturbed by PDT; however, retinal sensitivity in the macular area of eyes with subfoveal PCV improved shortly after PDT (70).
Intravitreal bevacizumab or ranibizumab for PCV
Although PCV appears likely to be refractory to anti-VEGF therapy, the exact role of VEGF in the pathogenesis of PCV remains to be elucidated. The first findings in surgical specimens of PCV demonstrated positive immunohistochemical staining for VEGF in the RPE and vascular endothelial cells. However, recent studies showed negative immunostaining in vascular cells for VEGF (71).
In another study, Tong et al. (72) found that the VEGF levels in aqueous humor in PCV were significantly lower than those in neovascular AMD.
Moreover, recent studies analyzing PCV lesions refractory to anti-VEGF therapy suggest that the development of PCV is less likely to be dependent on a VEGF-related pathway than is more typical type 1 neovascularization (73).
A recent report showed that ranibizumab leads to stabilization of vision, resolution of subretinal hemorrhage, and decrease in the PCV-associated macular edema. However, polypoidal complexes decreased in only 33% of the patients (76).
In the EVEREST study ranibizumab monotherapy group, the visual acuity improved in spite of the presence of polyps. This finding may be attributed to the antipermeability effects of ranibizumab, thereby resolving retinal thickening and exudate accumulation. These findings are consistent with those reported earlier with other anti-VEGF therapies.
Cho et al. (77) compared the effectiveness of intravitreal injection of bevacizumab and ranibizumab in 121 patients with PCV. Patients received three or more injections of 1.25 mg intravitreal bevacizumab, or 0.5 mg intravitreal ranibizumab, and outcomes were assessed after 12 months. They found no significant differences between both study groups in BCVA outcomes or central foveal thickness improvement. Polyp regression was also similar in both groups, with a regression rate of 24% in bevacizumab patients and 23% in the ranibizumab group.
There are no currently published studies comparing aflibercept (Eylea) with ranibizumab or bevacizumab.
It is well known that PDT induced up-regulation of VEGF. Therefore, treatment with both PDT, to resolve the polypoidal lesions, and anti-VEGF agents, to control up-regulation of VEGF, may be a reasonable strategy to maintain or improve the visual acuity and the anatomic changes in patients with PCV.
In the EVEREST study, the results showed the superior efficacy of intravitreal ranibizumab and PDT in achieving complete regression of polypoidal lesions (primary outcome), and BCVA (secondary outcomes).
Saito et al. (78) reported the efficacy of ranibizumab combined with PDT for patients with PCV for improving visual acuity and decreasing retinal and choroidal thickness without adverse events over 12 months. Recently, Ruamviboonsuk et al. (79) also reported the efficacy of combination therapy in 12 patients with PCV.
The ideal interval between intravitreal injection of anti-VEGF agents and PDT is controversial. Saito and collaborators administered PDT 1 or 2 days after intravitreal ranibizumab injection. Sato et al. (80) administered PDT 7 days after intravitreal bevacizumab, whereas Gomi et al. (81) administered PDT 1 day after intravitreal bevacizumab.
Currently, several groups have started to investigate combination therapy with PDT.
ASSESSMENTS AND RETREATMENT
Efficacy assessments include the evaluation of anatomical endpoints (ICGA-assessed polyp regression based on polyp area and total lesion area, OCT-measured CRT, leakage assessed by FA) and functional changes (assessed BCVA).
In the EVEREST Study, ICGA was performed at baseline and at months 3, 4, 5, and 6 to evaluate the polyp area. The researchers measured the area of the best-fit circles around all individual polyps excluding the hypofluorescent halo.
Patients were placed into one of four groups based on the polyp regression seen on ICGA: (i) Complete regression: no polyps seen on the imaging; (ii) partial regression: >10% decrease in polyp area compared with screening; (iii) no change: ≤10% change of polyp area compared with screening; and (iv) increase: >10% increase in polyp area compared with screening.
Immediately after the completion of ICGA, FA was performed through dilated pupils to assess presence of leakage and to obtain fundus photographs at baseline and at months 3, 4, 5, and 6. Changes in CRT were determined by OCT measurements at baseline and at months 1 to 6 (33). Patients with partial or no polyp regression after the initial 3 months were retreated with their assigned therapy. Patients with complete polyp regression by ICGA and leakage on FA were retreated with verteporfin PDT (or sham), and those with a decrease in BCVA of ≥5 letters were retreated with ranibizumab (or sham) (33).
We would suggest a protocol similar to that of Saito et al. (78) of follow-up examinations including evaluation of OCT every month, FA and ICGA every 3 months until the polypoidal lesions show complete regression, and then treatment restart if new exudative changes or subretinal hemorrhages are seen on fundus examination or OCT (Fig. 7.5).
Figure 7.5 Graph shows retreatment criteria after month tree treatment. Follow-up examinations included evaluation of the OCT images every month. (IVR, intravitreal ranibizumab injection. IVR + PDT, intravitreal ranibizumab injection plus PDT; PRN, pro re nata.) (Reprinted from Saito M, Iida T, Kano M. Combined intravitreal ranibizumab and photodynamic therapy for polypoidal choroidal vasculopathy. Retina. 2012;32(7):1272–1279. doi: 10.1097/IAE.0b013e318236e624, with permission.)