- Email: [email protected]
Identification of ingrowth site of idiopathic subfoveal choroidal neovascularization by indocyanine green angiography
Identification of Ingrowth Site of Idiopathic Subfoveal Choroidal Neovascularization by Indocyanine Green Angiography Fumio Shiraga, MD, Chieko Shiragami, MD, Toshihiko Matsuo, MD, Shiho Yokoe, MD, Ippei Takasu, MD, Hiroshi Ohtsuki, MD Purpose: This study aimed to determine whether indocyanine green (ICG) angiography is useful to identify the ingrowth site of idiopathic choroidal neovascularization (CNV), which can predict visual outcomes after surgical removal of idiopathic CNV. Design: Consecutive, observational case series. Participants: Twenty-six patients with idiopathic subfoveal CNV, of whom six underwent submacular surgery. Intervention: Indocyanine green videoangiography with a scanning laser ophthalmoscope. Main Outcome Measures: We studied ICG videoangiographic images of choroidal neovascular membranes from the early phase to the late phase with special attention to abnormal findings, which can indicate the ingrowth site of CNV. Results: Early ICG angiography demonstrated distinct neovascular vessels in 24 of the 26 patients (92%). Hypofluorescent rims continuously or intermittently surrounded neovascular membranes on late ICG angiograms in 21 of the 26 patients (81%). In 22 of the 26 patients (85%), ICG angiography demonstrated hypofluorescent areas within the CNV. These hypofluorescent areas frequently became ring shaped in the middle to late phase of the ICG angiography. In 14 of 16 patients (88%) with CNV larger than half a disc area, the filling of neovascular vessels appeared from the inside of the hypofluorescent areas and branched out toward the surrounding hyperfluorescent membrane in the early phase. In all six patients who underwent surgical removal of CNV, ICG videoangiography showed these hypofluorescent areas from which neovascular vessels emanated. Three of the four surgical patients, in whom hypofluorescent areas or central fluorescent areas surrounded by ring-shaped hypofluorescence were extrafoveal or juxtafoveal, had a best postoperative visual acuity of 20/60 or better. In contrast, both surgical patients with subfoveal hypofluorescent areas had a best postoperative visual acuity of 20/70 or worse. Conclusions: Although further observations are needed, ICG angiography may be a useful adjunct in the identification of the ingrowth site of idiopathic CNV, which can predict visual outcomes after surgery. Ophthalmology 2000;107:600 – 607 © 2000 by the American Academy of Ophthalmology. Although improvement in vitreous surgical techniques has enabled removal of subretinal neovascular membranes, the visual outcomes after surgery depend on the underlying disease.1– 6 The removal of subfoveal membranes does not improve central vision in most patients with age-related macular degeneration (AMD).2,3 In contrast, this surgical treatment may result in excellent recovery of good vision in Originally received: December 9, 1998. Accepted: November 2, 1999. Manuscript no. 98786. Department of Ophthalmology, Okayama University Medical School, Okayama, Japan. Presented in part as a poster at the annual meeting of the American Academy of Ophthalmology, San Francisco, California, October 1997. Supported in part by Health Sciences Research grants from the Ministry of Health and Welfare, Tokyo, Japan. The authors have no proprietary interest in any materials used in this study. Reprint requests to Fumio Shiraga, MD, Department of Ophthalmology, Okayama University Medical School, 2-5-1 Shikata-cho, Okayama 7008558, Japan. E-mail: [email protected].
600
© 2000 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
younger patients with presumed ocular histoplasmosis syndrome (POHS) or idiopathic choroidal neovascularization.1,3,5,6 This discrepancy can be explained mainly by the difference in growth patterns of choroidal neovascularization (CNV). Gass7 classified CNV into two different histologic types, according to its location. Type 1 CNV is located beneath the retinal pigment epithelium (RPE), and some portion of the CNV enters the subretinal space as the proliferation continues.8 Most older patients with AMD experience type 1 CNV, which cannot be excised without removing the RPE at the fovea. Type 2 CNV, located between the sensory retina and the RPE, tends to develop in younger patients with POHS or idiopathic CNV. The submacular excision of type 2 CNV holds the possibility of sparing the RPE at the fovea with resulting useful postoperative vision. The differentiation between type 1 and type 2 CNV may be clinically possible with biomicroscopic slit-lamp examination with a contact lens, fluorescein angiography, and other information such as patient age or fundus appearance ISSN 0161-6420/00/$–see front matter PII S0161-6420(99)00126-8
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG Table 1. Preoperative and Postoperative Data in Patients with Idiopathic Subfoveal Choroidal Neovascularization Patient No.
Age (yrs)/ Gender
Preoperative
Postoperative
1 2 3 4 5 6
42/F 47/F 36/M 49/F 39/M 48/M
20/400 20/200 20/500 20/200 20/1000 20/400
20/60 20/50 20/100 20/70 20/60 20/200
Final
Pigmented Halo or Plaque
Focal Hyperfluorescence Areas (FA)
20/60 20/50 20/100 20/100 20/1000 20/360
⫹ ⫺ ⫹ ⫾ ⫹ ⫺
juxtafoveal — extrafoveal subfoveal extrafoveal —
Best Visual Acuity
Hypofluorescence No. Areas (ICG)* Recurrences juxtafoveal extrafoveal juxtafoveal subfoveal extrafoveal subfoveal
Followup (mos)
1 0 0 1 0 0
63 60 52 46 18 17
FA ⫽ fluorescein angiography; ICG ⫽ indocyanine green angiography. * In eyes with ring-shaped hypofluorescence, the location of central fluorescent areas surrounded by the ring-shaped hypofluorescence is shown.
in fellow eyes.7,9 Although type 2 CNV may be eligible for surgical excision, a defect of the subfoveal RPE is inevitable in patients with a subfoveal ingrowth site of CNV. In patients with an extrafoveal ingrowth site, the subfoveal RPE may remain intact after the surgery, resulting in possible recovery of visual acuity. Melberg et al10 reported that the ingrowth site of subfoveal CNV, using color fundus photography and fluorescein angiography, was identified in 60 of 84 eyes (71%) with subfoveal CNV associated with POHS, multifocal choroiditis, and idiopathic CNV, and could predict visual outcomes after surgical removal of CNV. However, additional information is desired to increase the identification rate of the ingrowth site. Indocyanine green (ICG) angiography can provide images of choroidal neovascular membranes through blood or serous fluid. In addition, ICG videoangiography, with a scanning laser ophthalmoscope, enables the visualization of dynamics of the early transit from the filling of choroidal artery to the filling of neovascular nets.11 However, the identification of the ingrowth site of idiopathic CNV by ICG angiography thus far has not been reported. Thus we studied ICG angiographic features of idiopathic CNV to determine the significance of ICG videoangiography in the identification of the ingrowth site of idiopathic CNV.
Patients and Methods From August 1992 through September 1997, we examined 26 consecutive patients with unilateral idiopathic subfoveal CNV who met the following criteria: (1) less than 50 years of age; (2) the presence of exudative manifestations, hemorrhagic manifestations, or both associated with subfoveal CNV; (3) the absence of drusen, retinal pigment epithelial detachment, angioid streaks, pathologic myopia (⫺6.0 diopters [D] or more), traumatic choroidal rupture, peripapillary changes with atrophic or pigmented “punched out” chorioretinal lesions, and uveitis; (4) no previous laser treatment; (5) no allergy to fluorescein, ICG, or iodine; and (6) signed informed consent. All patients were Japanese; 16 patients were female and 10 were male. The mean age was 39.7 years with a range of 27 to 49 years. Examinations at the initial visit included slit-lamp biomicroscopic fundus examination, fluorescein angiography, and ICG videoangiography. Indocyanine green videoangiography was performed with a scanning laser ophthalmoscope (Rodenstock Instrument, Munchen, Germany). Indocyanine green dye of 25 mg
dissolved in 1 ml water was pushed into the cubital vein by a flash of 20 ml saline to achieve high-resolution images in the early phase. We used a 20° field size in the early arterial phase, and regular angiography, using a 40° field size, was performed later. After regular angiography for 20 to 30 minutes, we performed a landmark injection of ICG dye of 12.5 mg, showing retinal vascular landmark. We studied the ICG videoangiographic images of CNV from the early phase to the late phase, with regard to identification of neovascular vessels, a hypofluorescent rim at the margin of the CNV, and a hypofluorescent area within the CNV. In addition, we reviewed color fundus photographs and fluorescein angiograms in an attempt to identify a pigmented halo or plaque in neovascular complexes, which occurs around the site of origin of the CNV.7 Surgical removal of subfoveal neovascular membrane was performed in 6 of the 26 patients. All six patients fulfilled our surgical eligibility: (1) sharply defined borders and plaque-like elevation of subfoveal neovascular membranes, (2) subfoveal CNV larger than 1 disc area, (3) fluorescein angiography demonstrating classic CNV, (4) preoperative visual acuity of 20/100 or worse, and (5) signed informed consent. The reading of preoperative fundus photographs and ICG angiograms was performed by one of the authors (SY) in a masked fashion without knowledge of surgical outcome. We studied the correlation between the location of the hypofluorescent area on preoperative ICG angiograms and the best postoperative visual acuity. In patient 4 (Table 1), histopathologic examination of a surgically excised membrane was performed. The surgeon attempted to remove the membrane intact. The orientation of the membrane was determined by the landmark of remarkable pigmentation seen on the external side (choroid side) of the membrane. The specimen was fixed in 10% neutral buffered formalin and processed for routine light microscopic examination.
Results Indocyanine Green Angiographic Findings of Choroidal Neovascularization Neovascular Vessels. Early ICG angiography, with a 20° field size, showed neovascular vessels as distinctly visible as the retinal vessels in 24 of 26 patients (92%). The filling of most of these neovascular vessels appeared before the filling of the retinal vessels and branched out toward the surrounding neovascular membrane (Fig 1). These vessels were visible several minutes after the injection of ICG dye and were masked by the staining of neovascular membranes. Hypofluorescent Rim. In 21 of 26 patients (81%), hypofluorescent lines, independent of subretinal blood or exudates, demar-
601
Ophthalmology Volume 107, Number 3, March 2000
Figure 1. An early indocyanine green angiogram with a 20° field size demonstrates distinct neovascular vessels emanating from the inside of a hypofluorescent area.
cated neovascular membranes from the surrounding choroidal tissues (Figs 2, 3 and 4). The hypofluorescent lines, continuously or intermittently, surrounded the whole neovascular membrane. These rims of neovascular membranes often became increasingly hypofluorescent in the later phases of the ICG angiography and
appeared to be more evident in patients with longer duration from the onset. Hypofluorescent Area. In 22 of 26 patients (85%), early to late ICG angiography demonstrated a hypofluorescent area within CNV, independent of subretinal blood or exudates (Fig 2). This hypofluorescence was seen in all 16 patients with CNV larger than half a disc area. In 14 of these 16 patients (88%), the filling of neovascular vessels appeared from the inside of this hypofluorescent area and branched out toward the surrounding hyperfluorescent membrane (Figs 1 and 2). In the middle to late phase of the angiography, the hypofluorescence frequently became ring shaped because of the continuance of the fluorescence in the central area from which neovascular vessels emanated. These findings suggest that the ingrowth site of CNV, which can predict whether the subfoveal RPE remains preserved after surgical removal of neovascular membranes, may be located within the hypofluorescent area with central fluorescence. The hypofluorescent area varied in size from 0.04 to 1 disc area. In 16 of the 22 patients (73%), the hypofluorescent area or the central fluorescent area in eyes with ring-shaped hypofluorescence did not involve the fovea.
Correlation between Location of the Hypofluorescent Area and Visual Outcomes after Surgical Removal of Neovascular Membranes The patient data is shown in Table 1. Mean follow-up period was 42.7 months with a range of 17 to 63 months.
Figure 2. A, color photograph demonstrates a subfoveal lesion surrounded intermittently by a subretinal hemorrhage. B, on fluorescein angiography, a bright hyperfluorescent area is seen that can demonstrate a classic choroidal neovascularization. C,D, early (C) and late (D) indocyanine green angiography demonstrates a hypofluorescent area (surrounded by arrows). In the early phase (C), neovascular vessels appeared from the inside of this hypofluorescent area (arrowhead) and branched out toward the margin of the membrane. D, a hypofluorescent rim partially corresponds to discontinuous subretinal blood seen in color photograph (A). However, this rim is more continuous than the subretinal blood, especially in the temporal margin of the membrane.
602
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG
Figure 3. A 49-year-old patient who underwent removal of choroidal neovascularization (patient 4). A, a preoperative color photograph shows a well-defined subfoveal neovascular membrane. B, early indocyanine green (ICG) angiography (with a 20° field size) demonstrates distinct neovascular vessels and a hypofluorescent rim. C, late ICG angiography (with a 20° field size) demonstrates a hypofluorescent area. D, a postoperative color photograph 2 months after the surgery. The defect of retinal pigment epithelium and Bruch’s membrane involves the fovea. E, the hypofluorescence area remains in late ICG angiography (with a 20° field size), but it is smaller than before the surgery. The hypofluorescent rim disappeared. F, an excised specimen of the neovascular membrane shows reactive retinal pigment epithelial (RPE) cells on the external aspect of the membrane (arrowheads; stain, periodic acid-Schiff; original magnification ⫻20). The study of serial sections showed that there was remarkable proliferation of RPE cells (arrow) in the hypofluorescent area.
603
Ophthalmology Volume 107, Number 3, March 2000
Figure 4. A 42-year-old patient who underwent surgical removal of choroidal neovascularization (CNV; patient 1). A, a fluorescein angiogram at the onset (1 year before the surgery), which had been taken at an affiliated hospital, demonstrates juxtafoveal CNV. B, a preoperative color photograph shows a subfoveal lesion shaped as a CNV. A discontinuous pigmented halo is seen. C, a preoperative fluorescein angiogram demonstrates a classic pattern. D,E, indocyanine green (ICG) angiography shows neovascular vessels (arrows in D), a hypofluorescent rim (arrows in E), and discontinuous ring-shaped hypofluorescence (arrowheads in E). The fovea is located within the temporal area (arrowhead) of the ring-shaped hypofluorescence on the early ICG angiogram (D). The location of the central fluorescent area surrounded by the discontinuous ring-shaped hypofluorescence corresponds to the initial site of the CNV seen at the onset. F, after the surgery, the subfoveal retinal pigment epithelium appears to remain. Persistence of CNV is seen (arrow). An endolaser scar around the retinotomy is present superior to the macula. G, on a postoperative ICG angiogram, the hypofluorescent rim and the ring-shaped hypofluorescence are not apparent. An arrow indicates hypofluorescence corresponding to the endolaser scar.
604
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG The hypofluorescent area on preoperative ICG angiograms was identified in all six patients. In two patients, the fovea was located within the hypofluorescent area or the central fluorescent area (Fig 3). In the remaining four patients, the fovea was present outside the hypofluorescent area or within the area of ring-shaped hypofluorescence (Fig 4). However, it was possible to locate pigmented halos or plaques, which occur around the site of origin of the CNV, in three of the six patients. In patient 4 (Table 1), the pigmented lesion was too faint to locate. The early frame of fluorescein angiograms, in four of the six patients, demonstrated areas of focal hyperfluorescence from which the membrane appeared to arise. In the remaining two patients, the membrane first appeared as a diffuse area of hyperfluorescence without a focal origin. Five of six patients had a best postoperative visual acuity of 20/100 or better. In two of these five patients, however, the visual acuity decreased after the attainment of the best postoperative visual acuity as a result of recurrence of CNV or development of retinal detachment. Three of the four patients, in whom hypofluorescent areas or central fluorescent areas surrounded by ringshaped hypofluorescence were extrafoveal or juxtafoveal, had a best postoperative visual acuity of 20/60 or better. In contrast, both patients with subfoveal hypofluorescent areas had a best postoperative visual acuity of 20/70 or worse. In two patients with an extrafoveal hypofluorescent area and one patient with a juxtafoveal central fluorescent area, the subfoveal RPE seemed to remain preserved after the surgery (Fig 4F). In the remaining patient with a juxtafoveal CNV (patient 3 in Table 1), it was not evident whether subfoveal RPE was absent or atrophic. The subfoveal RPE defect after the surgery was evident in both patients who had a subfoveal hypofluorescent area (Fig 3D). The neovascular vessels and the hypofluorescent rim, seen on preoperative ICG angiograms, disappeared after the surgery (Figs 3E and 4G). In contrast, the hypofluorescent area did not disappear, and its size either remained unchanged or decreased after the surgery (Figs 3E and 4G).
Discussion Idiopathic CNV appears to be a distinct subset found in young patients in whom CNV develops in the absence of other underlying ocular pathologic conditions. Although many reports12–17 showed ICG angiographic findings of CNV, most of the patients had AMD. Recently, Iida et al17 reported a dark rim surrounding idiopathic CNV and choroidal vascular abnormalities on ICG angiography. However, no report has yet shown abnormal ICG angiographic findings indicating the ingrowth site of idiopathic CNV, which may predict postoperative visual outcomes. Even with ICG angiography, neovascular vessels posterior to the RPE are often difficult to identify because of hemorrhage or exudates between the sensory retina and RPE or between the RPE and Bruch’s membrane. In contrast, neovascular vessels anterior to the RPE are visible as distinctly as retinal vessels on early ICG angiograms. In patients with AMD who have a “mixed” pattern of CNV,9 early ICG angiography often demonstrates distinct neovascular vessels, which are supposed to develop into the subretinal space beyond the RPE. Fluorescein angiography also enables the visualization of neovascular vessels anterior to the RPE. In idiopathic CNV, fluorescein angiography demonstrates a classic pattern; thus ICG angiography may not be needed to determine the extent and location of the CNV.
However, ICG angiography, which is less subject to blockage by subretinal hemorrhage or fluid, may reveal these neovascular vessels more frequently and distinctly. In our study, neovascular vessels were identified in 92% of the patients by using ICG videoangiography with a scanning laser ophthalmoscope, whereas distinct neovascular vessels were visible on fluorescein angiograms in 17 of 26 patients (65%). In addition, the ICG angiography, with a 20° field size, demonstrated the early transit from the appearance of an ingrowth stalk to the filling of new vessels at the margin of neovascular complex. A hypofluorescent rim, independent of subretinal blood or exudates, surrounding neovascular membranes was seen in 81% of our patients with idiopathic CNV. This finding was also seen in patients with CNV secondary to AMD.14 The hypofluorescent rim of type 1 CNV corresponds to elevated blocked fluorescence on fluorescein angiograms resulting from blockage of the underlying choroidal fluorescence or neovascular channel leakage by hyperplastic RPE, hypertrophied RPE, or both.18 However, the hypofluorescent rim of idiopathic CNV may be mainly the result of the blockage by the reactive RPE cells on the external aspect of neovascular membranes (Fig 3F), which are histologically seen in type 2 CNV.7,9 In our six patients who underwent surgical excision of neovascular membranes, the hypofluorescent rim disappeared after surgery. Iida et al17 noted that the dark rim of idiopathic CNV on ICG angiograms became prominent during the follow-up period. This fact supports that CNV may be surrounded by reactive RPE cells in the process toward regression. It is possible to identify the ingrowth site of CNV, which predicts visual outcomes after surgical removal of neovascular membranes, even without ICG angiography. As mentioned above, Melberg et al10 reported that the ingrowth site was identifiable before surgery in 60 of 84 eyes (71%) with POHS (67 eyes), multifocal choroiditis (9 eyes), or idiopathic CNV (8 eyes) using color fundus photographs and fluorescein angiograms. As described in this report, the early phase of fluorescein angiography can demonstrate a visible stalk or an area of focal hyperfluorescence from which the membrane appeared to arise. A lighter colored spot within CNV frequently indicates the ingrowth site of CNV, especially in eyes with POHS in which neovascular buds originate from choroidal vessels within chorioretinal scars.7 In our study, the rate of identification by pigmented lesions on color photographs was lower (54%), and the pigmentation often appeared to be too discontinuous to identify the ingrowth site of CNV. Light-colored spots on color photographs or focal ingrowth stalks on fluorescein angiograms, which Melberg et al10 focused on, may be more predictable than pigmented lesions. Although the sample size of the present study is small, in contrast, hypofluorescent areas on ICG angiograms, including ring-shaped hypofluorescence, were identified in 85% of patients with idiopathic CNV. Moreover, these hypofluorescent areas were identifiable in all 16 patients with CNV larger than half a disc area, which may be more amenable to the surgical intervention. Although only six patients in our study group underwent surgical removal of CNV, in five of the six patients the location of the hypofluorescent area or
605
Ophthalmology Volume 107, Number 3, March 2000 the central hyperfluorescent area appeared to correspond to the site of the postoperative defect of the RPE, influencing postoperative visual outcomes. However, these hypofluorescent areas, including the ring-shaped hypofluorescence, tended to be ill-defined as compared with focal hyperfluorescent areas on the early phase of fluorescein angiography, which were identified in four of the six patients. Fluorescein angiography may locate the ingrowth sites more precisely than ICG angiography. Because the hypofluorescent area tended to decrease after surgery, it may be partially because of the blockage by reactive RPE cells (Fig 3F) surrounding the ingrowth stalk.7 In some cases, also in the early phase of fluorescein angiography, ring-shaped hypofluorescence surrounded focal hyperfluorescent areas from which the membrane appeared to arise, although it was afterward masked by leakage from CNV. Because ICG dye fluorescence is in the infrared spectrum and the image of the neovascular complex is less susceptible to masking by leakage from CNV or blockage by hemorrhage or exudates, ICG angiography could distinctly visualize hypofluorescence as a result of reactive RPE around the ingrowth site. In the late phase of ICG angiography, the intensity of the fluorescence in the central area may be determined by staining of or leakage from neovascular vessels and blockage by fibrous tissue or reactive RPE cells within membranes. Because damaged RPE can cause choriocapillary atrophy,19 defects in the RPE or Bruch’s membrane at the site of origin of the neovascularization may lead to the choriocapillary atrophy manifesting as hypofluorescence in ICG angiography. Although we have not performed ICG angiography on patients with other underlying pathologic conditions, such as POHS or multifocal choroiditis, we suggest that ICG angiography could also be a useful adjunct in the identification of the ingrowth site in these entities as well. To summarize the ICG findings of idiopathic CNV in our study, early ICG angiography detected distinct neovascular vessels emanating from the inside of a hypofluorescent area within CNV and afterward branching out toward the margin of the CNV. The later angiography demonstrated a hypofluorescent rim surrounding the CNV as well as the continuance of the hypofluorescent area often becoming ring shaped. These findings appear to be different from ICG findings in patients with AMD and may be characteristic of type 2 CNV. Even though biomicroscopic fundus examination and other information (patient age or fundus appearance of fellow eyes) enable the differentiation between type 1 and type 2,7,9 ICG angiography may be of assistance in the precise differentiation between them. Although the natural history of idiopathic CNV may be relatively favorable, patients with large-sized CNV may have a poor visual prognosis.20 The Macular Photocoagulation Study Group21,22 proved the beneficial effect of laser treatment in extrafoveal or juxtafoveal idiopathic CNV. However, there was no information on the laser treatment of subfoveal idiopathic CNV. Thomas et al3 reported that four of eight patients had a final visual acuity of 20/100 or better after surgical removal of subfoveal neovascular membranes in idiopathic CNV. Four of six patients in our study had a final visual acuity of 20/100 or better after surgery. Because
606
idiopathic CNV appears to be focal in nature and represents a type 2 configuration, selected patients with this disease may be amenable to surgical intervention. Grossniklaus et al9 noted that type 2 membranes may have extended from an extramacular source of CNV under the macula and that removal of membranes with this configuration may lead to an excellent visual outcome. In the present study, hypofluorescent areas or fluorescent areas in eyes with ring-shaped hypofluorescence did not involve the fovea in 16 of the 22 patients (73%) with hypofluorescent areas. The limitation of the present study is the small sample size. As described above, biomicroscopic fundus examination and fluorescein angiography are useful means in identification of the ingrowth site of CNV. Given the cost, ICG angiography should be performed only on patients in whom the ingrowth site cannot be identified by biomicroscopic fundus examination and fluorescein angiography. Currently, the Submacular Surgery Trial is evaluating the safety and efficacy of surgery for POHS and idiopathic CNV; should a benefit be proven, ICG angiography may be a useful adjunct in the identification of the ingrowth site of idiopathic CNV, although further investigation is needed.
References 1. Thomas MA, Kaplan HJ. Surgical removal of subfoveal neovascularization in the presumed ocular histoplasmosis syndrome. Am J Ophthalmol 1991;111:1–7. 2. Lambert HM, Capone A Jr, Aaberg TM, et al. Surgical excision of subfoveal neovascular membranes in age-related macular degeneration. Am J Ophthalmol 1992;113: 257– 62. 3. Thomas MA, Dickinson JD, Melberg NS, et al. Visual results after surgical removal of subfoveal choroidal neovascular membranes. Ophthalmology 1994;101:1384 –96. 4. Adelberg DA, Del Priore LV, Kaplan HJ. Surgery for subfoveal membranes in myopia, angioid streaks, and other disorders. Retina 1995;15:198 –205. 5. Holekamp NM, Thomas MA, Dickinson JD, Valluri S. Surgical removal of subfoveal choroidal neovascularization in presumed ocular histoplasmosis: stability of early visual results. Ophthalmology 1997;104:22– 6. 6. Olsen TW, Capone A Jr, Sternberg P Jr, et al. Subfoveal choroidal neovascularization in punctate inner choroidopathy. Surgical management and pathologic findings. Ophthalmology 1996;103:2061–9. 7. Gass JDM. Biomicroscopic and histopathologic considerations regarding the feasibility of surgical excision of subfoveal neovascular membranes. Am J Ophthalmol 1994;118: 285–98. 8. Green WR, Enger C. Age-related macular degeneration histopathologic studies. The 1992 Lorenz E. Zimmerman Lecture. Ophthalmology 1993;100:1519 –35. 9. Grossniklaus HE, Gass JDM. Clinicopathologic correlations of surgically excised type 1 and type 2 submacular choroidal neovascular membranes. Am J Ophthalmol 1998;126:59 – 69. 10. Melberg NS, Thomas MA, Burgess DB. The surgical removal of subfoveal choroidal neovascularization. Ingrowth site as a predictor of visual outcome. Retina 1996;16:190 –5. 11. Shiraga F, Ojima Y, Matsuo T, et al. Feeder vessel photocoagulation of subfoveal choroidal neovascularization secondary
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG
12. 13. 14.
15.
16.
to age-related macular degeneration. Ophthalmology 1998; 105:662–9. Destro M, Puliafito CA. Indocyanine green videoangiography of choroidal neovascularization. Ophthalmology 1989;96: 846 –53. Yannuzzi LA, Slakter JS, Sorenson JA, et al. Digital indocyanine green videoangiography and choroidal neovascularization. Retina 1992;12:191–223. Scheider A, Kaboth A, Neuhauser L. Detection of subretinal neovascular membranes with indocyanine green and an infrared scanning laser ophthalmoscope. Am J Ophthalmol 1992; 113:45–51. Chang TS, Freund KB, de la Cruz Z, et al. Clinicopathologic correlation of choroidal neovascularization demonstrated by indocyanine green angiography in a patient with retention of good vision for almost four years. Retina 1994;14:114 –24. Lee BL, Lim JI, Grossniklaus HE. Clinicopathologic features of indocyanine green angiography-imaged, surgically excised choroidal neovascular membranes. Retina 1996;16:64 –9.
17. Iida T, Hagimura N, Kishi S, Shimizu K. Indocyanine green angiographic features of idiopathic submacular choroidal neovascularization. Am J Ophthalmol 1998;126:70 – 6. 18. Lopez PF, Lambert HM, Grossniklaus HE, Sternberg P Jr. Well-defined subfoveal choroidal neovascular membranes in age-related macular degeneration. Ophthalmology 1993;100: 415–22. 19. Korte GE, Reppucci V, Henkind P. RPE destruction causes choriocapillary atrophy. Invest Ophthalmol Vis Sci 1984;25: 1135– 45. 20. Ho AC, Yannuzzi LA, Pisicano K, DeRosa J. The natural history of idiopathic subfoveal choroidal neovascularization. Ophthalmology 1995;102:782–9. 21. Argon laser photocoagulation for idiopathic neovascularization. Results of a randomized clinical trial. Arch Ophthalmol 1983;101:1358 – 61. 22. Krypton laser photocoagulation for idiopathic neovascular lesions. Results of a randomized clinical trial. Macular Photocoagulation Study Group. Arch Ophthalmol 1990;108: 832–7.
607
600
© 2000 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
younger patients with presumed ocular histoplasmosis syndrome (POHS) or idiopathic choroidal neovascularization.1,3,5,6 This discrepancy can be explained mainly by the difference in growth patterns of choroidal neovascularization (CNV). Gass7 classified CNV into two different histologic types, according to its location. Type 1 CNV is located beneath the retinal pigment epithelium (RPE), and some portion of the CNV enters the subretinal space as the proliferation continues.8 Most older patients with AMD experience type 1 CNV, which cannot be excised without removing the RPE at the fovea. Type 2 CNV, located between the sensory retina and the RPE, tends to develop in younger patients with POHS or idiopathic CNV. The submacular excision of type 2 CNV holds the possibility of sparing the RPE at the fovea with resulting useful postoperative vision. The differentiation between type 1 and type 2 CNV may be clinically possible with biomicroscopic slit-lamp examination with a contact lens, fluorescein angiography, and other information such as patient age or fundus appearance ISSN 0161-6420/00/$–see front matter PII S0161-6420(99)00126-8
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG Table 1. Preoperative and Postoperative Data in Patients with Idiopathic Subfoveal Choroidal Neovascularization Patient No.
Age (yrs)/ Gender
Preoperative
Postoperative
1 2 3 4 5 6
42/F 47/F 36/M 49/F 39/M 48/M
20/400 20/200 20/500 20/200 20/1000 20/400
20/60 20/50 20/100 20/70 20/60 20/200
Final
Pigmented Halo or Plaque
Focal Hyperfluorescence Areas (FA)
20/60 20/50 20/100 20/100 20/1000 20/360
⫹ ⫺ ⫹ ⫾ ⫹ ⫺
juxtafoveal — extrafoveal subfoveal extrafoveal —
Best Visual Acuity
Hypofluorescence No. Areas (ICG)* Recurrences juxtafoveal extrafoveal juxtafoveal subfoveal extrafoveal subfoveal
Followup (mos)
1 0 0 1 0 0
63 60 52 46 18 17
FA ⫽ fluorescein angiography; ICG ⫽ indocyanine green angiography. * In eyes with ring-shaped hypofluorescence, the location of central fluorescent areas surrounded by the ring-shaped hypofluorescence is shown.
in fellow eyes.7,9 Although type 2 CNV may be eligible for surgical excision, a defect of the subfoveal RPE is inevitable in patients with a subfoveal ingrowth site of CNV. In patients with an extrafoveal ingrowth site, the subfoveal RPE may remain intact after the surgery, resulting in possible recovery of visual acuity. Melberg et al10 reported that the ingrowth site of subfoveal CNV, using color fundus photography and fluorescein angiography, was identified in 60 of 84 eyes (71%) with subfoveal CNV associated with POHS, multifocal choroiditis, and idiopathic CNV, and could predict visual outcomes after surgical removal of CNV. However, additional information is desired to increase the identification rate of the ingrowth site. Indocyanine green (ICG) angiography can provide images of choroidal neovascular membranes through blood or serous fluid. In addition, ICG videoangiography, with a scanning laser ophthalmoscope, enables the visualization of dynamics of the early transit from the filling of choroidal artery to the filling of neovascular nets.11 However, the identification of the ingrowth site of idiopathic CNV by ICG angiography thus far has not been reported. Thus we studied ICG angiographic features of idiopathic CNV to determine the significance of ICG videoangiography in the identification of the ingrowth site of idiopathic CNV.
Patients and Methods From August 1992 through September 1997, we examined 26 consecutive patients with unilateral idiopathic subfoveal CNV who met the following criteria: (1) less than 50 years of age; (2) the presence of exudative manifestations, hemorrhagic manifestations, or both associated with subfoveal CNV; (3) the absence of drusen, retinal pigment epithelial detachment, angioid streaks, pathologic myopia (⫺6.0 diopters [D] or more), traumatic choroidal rupture, peripapillary changes with atrophic or pigmented “punched out” chorioretinal lesions, and uveitis; (4) no previous laser treatment; (5) no allergy to fluorescein, ICG, or iodine; and (6) signed informed consent. All patients were Japanese; 16 patients were female and 10 were male. The mean age was 39.7 years with a range of 27 to 49 years. Examinations at the initial visit included slit-lamp biomicroscopic fundus examination, fluorescein angiography, and ICG videoangiography. Indocyanine green videoangiography was performed with a scanning laser ophthalmoscope (Rodenstock Instrument, Munchen, Germany). Indocyanine green dye of 25 mg
dissolved in 1 ml water was pushed into the cubital vein by a flash of 20 ml saline to achieve high-resolution images in the early phase. We used a 20° field size in the early arterial phase, and regular angiography, using a 40° field size, was performed later. After regular angiography for 20 to 30 minutes, we performed a landmark injection of ICG dye of 12.5 mg, showing retinal vascular landmark. We studied the ICG videoangiographic images of CNV from the early phase to the late phase, with regard to identification of neovascular vessels, a hypofluorescent rim at the margin of the CNV, and a hypofluorescent area within the CNV. In addition, we reviewed color fundus photographs and fluorescein angiograms in an attempt to identify a pigmented halo or plaque in neovascular complexes, which occurs around the site of origin of the CNV.7 Surgical removal of subfoveal neovascular membrane was performed in 6 of the 26 patients. All six patients fulfilled our surgical eligibility: (1) sharply defined borders and plaque-like elevation of subfoveal neovascular membranes, (2) subfoveal CNV larger than 1 disc area, (3) fluorescein angiography demonstrating classic CNV, (4) preoperative visual acuity of 20/100 or worse, and (5) signed informed consent. The reading of preoperative fundus photographs and ICG angiograms was performed by one of the authors (SY) in a masked fashion without knowledge of surgical outcome. We studied the correlation between the location of the hypofluorescent area on preoperative ICG angiograms and the best postoperative visual acuity. In patient 4 (Table 1), histopathologic examination of a surgically excised membrane was performed. The surgeon attempted to remove the membrane intact. The orientation of the membrane was determined by the landmark of remarkable pigmentation seen on the external side (choroid side) of the membrane. The specimen was fixed in 10% neutral buffered formalin and processed for routine light microscopic examination.
Results Indocyanine Green Angiographic Findings of Choroidal Neovascularization Neovascular Vessels. Early ICG angiography, with a 20° field size, showed neovascular vessels as distinctly visible as the retinal vessels in 24 of 26 patients (92%). The filling of most of these neovascular vessels appeared before the filling of the retinal vessels and branched out toward the surrounding neovascular membrane (Fig 1). These vessels were visible several minutes after the injection of ICG dye and were masked by the staining of neovascular membranes. Hypofluorescent Rim. In 21 of 26 patients (81%), hypofluorescent lines, independent of subretinal blood or exudates, demar-
601
Ophthalmology Volume 107, Number 3, March 2000
Figure 1. An early indocyanine green angiogram with a 20° field size demonstrates distinct neovascular vessels emanating from the inside of a hypofluorescent area.
cated neovascular membranes from the surrounding choroidal tissues (Figs 2, 3 and 4). The hypofluorescent lines, continuously or intermittently, surrounded the whole neovascular membrane. These rims of neovascular membranes often became increasingly hypofluorescent in the later phases of the ICG angiography and
appeared to be more evident in patients with longer duration from the onset. Hypofluorescent Area. In 22 of 26 patients (85%), early to late ICG angiography demonstrated a hypofluorescent area within CNV, independent of subretinal blood or exudates (Fig 2). This hypofluorescence was seen in all 16 patients with CNV larger than half a disc area. In 14 of these 16 patients (88%), the filling of neovascular vessels appeared from the inside of this hypofluorescent area and branched out toward the surrounding hyperfluorescent membrane (Figs 1 and 2). In the middle to late phase of the angiography, the hypofluorescence frequently became ring shaped because of the continuance of the fluorescence in the central area from which neovascular vessels emanated. These findings suggest that the ingrowth site of CNV, which can predict whether the subfoveal RPE remains preserved after surgical removal of neovascular membranes, may be located within the hypofluorescent area with central fluorescence. The hypofluorescent area varied in size from 0.04 to 1 disc area. In 16 of the 22 patients (73%), the hypofluorescent area or the central fluorescent area in eyes with ring-shaped hypofluorescence did not involve the fovea.
Correlation between Location of the Hypofluorescent Area and Visual Outcomes after Surgical Removal of Neovascular Membranes The patient data is shown in Table 1. Mean follow-up period was 42.7 months with a range of 17 to 63 months.
Figure 2. A, color photograph demonstrates a subfoveal lesion surrounded intermittently by a subretinal hemorrhage. B, on fluorescein angiography, a bright hyperfluorescent area is seen that can demonstrate a classic choroidal neovascularization. C,D, early (C) and late (D) indocyanine green angiography demonstrates a hypofluorescent area (surrounded by arrows). In the early phase (C), neovascular vessels appeared from the inside of this hypofluorescent area (arrowhead) and branched out toward the margin of the membrane. D, a hypofluorescent rim partially corresponds to discontinuous subretinal blood seen in color photograph (A). However, this rim is more continuous than the subretinal blood, especially in the temporal margin of the membrane.
602
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG
Figure 3. A 49-year-old patient who underwent removal of choroidal neovascularization (patient 4). A, a preoperative color photograph shows a well-defined subfoveal neovascular membrane. B, early indocyanine green (ICG) angiography (with a 20° field size) demonstrates distinct neovascular vessels and a hypofluorescent rim. C, late ICG angiography (with a 20° field size) demonstrates a hypofluorescent area. D, a postoperative color photograph 2 months after the surgery. The defect of retinal pigment epithelium and Bruch’s membrane involves the fovea. E, the hypofluorescence area remains in late ICG angiography (with a 20° field size), but it is smaller than before the surgery. The hypofluorescent rim disappeared. F, an excised specimen of the neovascular membrane shows reactive retinal pigment epithelial (RPE) cells on the external aspect of the membrane (arrowheads; stain, periodic acid-Schiff; original magnification ⫻20). The study of serial sections showed that there was remarkable proliferation of RPE cells (arrow) in the hypofluorescent area.
603
Ophthalmology Volume 107, Number 3, March 2000
Figure 4. A 42-year-old patient who underwent surgical removal of choroidal neovascularization (CNV; patient 1). A, a fluorescein angiogram at the onset (1 year before the surgery), which had been taken at an affiliated hospital, demonstrates juxtafoveal CNV. B, a preoperative color photograph shows a subfoveal lesion shaped as a CNV. A discontinuous pigmented halo is seen. C, a preoperative fluorescein angiogram demonstrates a classic pattern. D,E, indocyanine green (ICG) angiography shows neovascular vessels (arrows in D), a hypofluorescent rim (arrows in E), and discontinuous ring-shaped hypofluorescence (arrowheads in E). The fovea is located within the temporal area (arrowhead) of the ring-shaped hypofluorescence on the early ICG angiogram (D). The location of the central fluorescent area surrounded by the discontinuous ring-shaped hypofluorescence corresponds to the initial site of the CNV seen at the onset. F, after the surgery, the subfoveal retinal pigment epithelium appears to remain. Persistence of CNV is seen (arrow). An endolaser scar around the retinotomy is present superior to the macula. G, on a postoperative ICG angiogram, the hypofluorescent rim and the ring-shaped hypofluorescence are not apparent. An arrow indicates hypofluorescence corresponding to the endolaser scar.
604
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG The hypofluorescent area on preoperative ICG angiograms was identified in all six patients. In two patients, the fovea was located within the hypofluorescent area or the central fluorescent area (Fig 3). In the remaining four patients, the fovea was present outside the hypofluorescent area or within the area of ring-shaped hypofluorescence (Fig 4). However, it was possible to locate pigmented halos or plaques, which occur around the site of origin of the CNV, in three of the six patients. In patient 4 (Table 1), the pigmented lesion was too faint to locate. The early frame of fluorescein angiograms, in four of the six patients, demonstrated areas of focal hyperfluorescence from which the membrane appeared to arise. In the remaining two patients, the membrane first appeared as a diffuse area of hyperfluorescence without a focal origin. Five of six patients had a best postoperative visual acuity of 20/100 or better. In two of these five patients, however, the visual acuity decreased after the attainment of the best postoperative visual acuity as a result of recurrence of CNV or development of retinal detachment. Three of the four patients, in whom hypofluorescent areas or central fluorescent areas surrounded by ringshaped hypofluorescence were extrafoveal or juxtafoveal, had a best postoperative visual acuity of 20/60 or better. In contrast, both patients with subfoveal hypofluorescent areas had a best postoperative visual acuity of 20/70 or worse. In two patients with an extrafoveal hypofluorescent area and one patient with a juxtafoveal central fluorescent area, the subfoveal RPE seemed to remain preserved after the surgery (Fig 4F). In the remaining patient with a juxtafoveal CNV (patient 3 in Table 1), it was not evident whether subfoveal RPE was absent or atrophic. The subfoveal RPE defect after the surgery was evident in both patients who had a subfoveal hypofluorescent area (Fig 3D). The neovascular vessels and the hypofluorescent rim, seen on preoperative ICG angiograms, disappeared after the surgery (Figs 3E and 4G). In contrast, the hypofluorescent area did not disappear, and its size either remained unchanged or decreased after the surgery (Figs 3E and 4G).
Discussion Idiopathic CNV appears to be a distinct subset found in young patients in whom CNV develops in the absence of other underlying ocular pathologic conditions. Although many reports12–17 showed ICG angiographic findings of CNV, most of the patients had AMD. Recently, Iida et al17 reported a dark rim surrounding idiopathic CNV and choroidal vascular abnormalities on ICG angiography. However, no report has yet shown abnormal ICG angiographic findings indicating the ingrowth site of idiopathic CNV, which may predict postoperative visual outcomes. Even with ICG angiography, neovascular vessels posterior to the RPE are often difficult to identify because of hemorrhage or exudates between the sensory retina and RPE or between the RPE and Bruch’s membrane. In contrast, neovascular vessels anterior to the RPE are visible as distinctly as retinal vessels on early ICG angiograms. In patients with AMD who have a “mixed” pattern of CNV,9 early ICG angiography often demonstrates distinct neovascular vessels, which are supposed to develop into the subretinal space beyond the RPE. Fluorescein angiography also enables the visualization of neovascular vessels anterior to the RPE. In idiopathic CNV, fluorescein angiography demonstrates a classic pattern; thus ICG angiography may not be needed to determine the extent and location of the CNV.
However, ICG angiography, which is less subject to blockage by subretinal hemorrhage or fluid, may reveal these neovascular vessels more frequently and distinctly. In our study, neovascular vessels were identified in 92% of the patients by using ICG videoangiography with a scanning laser ophthalmoscope, whereas distinct neovascular vessels were visible on fluorescein angiograms in 17 of 26 patients (65%). In addition, the ICG angiography, with a 20° field size, demonstrated the early transit from the appearance of an ingrowth stalk to the filling of new vessels at the margin of neovascular complex. A hypofluorescent rim, independent of subretinal blood or exudates, surrounding neovascular membranes was seen in 81% of our patients with idiopathic CNV. This finding was also seen in patients with CNV secondary to AMD.14 The hypofluorescent rim of type 1 CNV corresponds to elevated blocked fluorescence on fluorescein angiograms resulting from blockage of the underlying choroidal fluorescence or neovascular channel leakage by hyperplastic RPE, hypertrophied RPE, or both.18 However, the hypofluorescent rim of idiopathic CNV may be mainly the result of the blockage by the reactive RPE cells on the external aspect of neovascular membranes (Fig 3F), which are histologically seen in type 2 CNV.7,9 In our six patients who underwent surgical excision of neovascular membranes, the hypofluorescent rim disappeared after surgery. Iida et al17 noted that the dark rim of idiopathic CNV on ICG angiograms became prominent during the follow-up period. This fact supports that CNV may be surrounded by reactive RPE cells in the process toward regression. It is possible to identify the ingrowth site of CNV, which predicts visual outcomes after surgical removal of neovascular membranes, even without ICG angiography. As mentioned above, Melberg et al10 reported that the ingrowth site was identifiable before surgery in 60 of 84 eyes (71%) with POHS (67 eyes), multifocal choroiditis (9 eyes), or idiopathic CNV (8 eyes) using color fundus photographs and fluorescein angiograms. As described in this report, the early phase of fluorescein angiography can demonstrate a visible stalk or an area of focal hyperfluorescence from which the membrane appeared to arise. A lighter colored spot within CNV frequently indicates the ingrowth site of CNV, especially in eyes with POHS in which neovascular buds originate from choroidal vessels within chorioretinal scars.7 In our study, the rate of identification by pigmented lesions on color photographs was lower (54%), and the pigmentation often appeared to be too discontinuous to identify the ingrowth site of CNV. Light-colored spots on color photographs or focal ingrowth stalks on fluorescein angiograms, which Melberg et al10 focused on, may be more predictable than pigmented lesions. Although the sample size of the present study is small, in contrast, hypofluorescent areas on ICG angiograms, including ring-shaped hypofluorescence, were identified in 85% of patients with idiopathic CNV. Moreover, these hypofluorescent areas were identifiable in all 16 patients with CNV larger than half a disc area, which may be more amenable to the surgical intervention. Although only six patients in our study group underwent surgical removal of CNV, in five of the six patients the location of the hypofluorescent area or
605
Ophthalmology Volume 107, Number 3, March 2000 the central hyperfluorescent area appeared to correspond to the site of the postoperative defect of the RPE, influencing postoperative visual outcomes. However, these hypofluorescent areas, including the ring-shaped hypofluorescence, tended to be ill-defined as compared with focal hyperfluorescent areas on the early phase of fluorescein angiography, which were identified in four of the six patients. Fluorescein angiography may locate the ingrowth sites more precisely than ICG angiography. Because the hypofluorescent area tended to decrease after surgery, it may be partially because of the blockage by reactive RPE cells (Fig 3F) surrounding the ingrowth stalk.7 In some cases, also in the early phase of fluorescein angiography, ring-shaped hypofluorescence surrounded focal hyperfluorescent areas from which the membrane appeared to arise, although it was afterward masked by leakage from CNV. Because ICG dye fluorescence is in the infrared spectrum and the image of the neovascular complex is less susceptible to masking by leakage from CNV or blockage by hemorrhage or exudates, ICG angiography could distinctly visualize hypofluorescence as a result of reactive RPE around the ingrowth site. In the late phase of ICG angiography, the intensity of the fluorescence in the central area may be determined by staining of or leakage from neovascular vessels and blockage by fibrous tissue or reactive RPE cells within membranes. Because damaged RPE can cause choriocapillary atrophy,19 defects in the RPE or Bruch’s membrane at the site of origin of the neovascularization may lead to the choriocapillary atrophy manifesting as hypofluorescence in ICG angiography. Although we have not performed ICG angiography on patients with other underlying pathologic conditions, such as POHS or multifocal choroiditis, we suggest that ICG angiography could also be a useful adjunct in the identification of the ingrowth site in these entities as well. To summarize the ICG findings of idiopathic CNV in our study, early ICG angiography detected distinct neovascular vessels emanating from the inside of a hypofluorescent area within CNV and afterward branching out toward the margin of the CNV. The later angiography demonstrated a hypofluorescent rim surrounding the CNV as well as the continuance of the hypofluorescent area often becoming ring shaped. These findings appear to be different from ICG findings in patients with AMD and may be characteristic of type 2 CNV. Even though biomicroscopic fundus examination and other information (patient age or fundus appearance of fellow eyes) enable the differentiation between type 1 and type 2,7,9 ICG angiography may be of assistance in the precise differentiation between them. Although the natural history of idiopathic CNV may be relatively favorable, patients with large-sized CNV may have a poor visual prognosis.20 The Macular Photocoagulation Study Group21,22 proved the beneficial effect of laser treatment in extrafoveal or juxtafoveal idiopathic CNV. However, there was no information on the laser treatment of subfoveal idiopathic CNV. Thomas et al3 reported that four of eight patients had a final visual acuity of 20/100 or better after surgical removal of subfoveal neovascular membranes in idiopathic CNV. Four of six patients in our study had a final visual acuity of 20/100 or better after surgery. Because
606
idiopathic CNV appears to be focal in nature and represents a type 2 configuration, selected patients with this disease may be amenable to surgical intervention. Grossniklaus et al9 noted that type 2 membranes may have extended from an extramacular source of CNV under the macula and that removal of membranes with this configuration may lead to an excellent visual outcome. In the present study, hypofluorescent areas or fluorescent areas in eyes with ring-shaped hypofluorescence did not involve the fovea in 16 of the 22 patients (73%) with hypofluorescent areas. The limitation of the present study is the small sample size. As described above, biomicroscopic fundus examination and fluorescein angiography are useful means in identification of the ingrowth site of CNV. Given the cost, ICG angiography should be performed only on patients in whom the ingrowth site cannot be identified by biomicroscopic fundus examination and fluorescein angiography. Currently, the Submacular Surgery Trial is evaluating the safety and efficacy of surgery for POHS and idiopathic CNV; should a benefit be proven, ICG angiography may be a useful adjunct in the identification of the ingrowth site of idiopathic CNV, although further investigation is needed.
References 1. Thomas MA, Kaplan HJ. Surgical removal of subfoveal neovascularization in the presumed ocular histoplasmosis syndrome. Am J Ophthalmol 1991;111:1–7. 2. Lambert HM, Capone A Jr, Aaberg TM, et al. Surgical excision of subfoveal neovascular membranes in age-related macular degeneration. Am J Ophthalmol 1992;113: 257– 62. 3. Thomas MA, Dickinson JD, Melberg NS, et al. Visual results after surgical removal of subfoveal choroidal neovascular membranes. Ophthalmology 1994;101:1384 –96. 4. Adelberg DA, Del Priore LV, Kaplan HJ. Surgery for subfoveal membranes in myopia, angioid streaks, and other disorders. Retina 1995;15:198 –205. 5. Holekamp NM, Thomas MA, Dickinson JD, Valluri S. Surgical removal of subfoveal choroidal neovascularization in presumed ocular histoplasmosis: stability of early visual results. Ophthalmology 1997;104:22– 6. 6. Olsen TW, Capone A Jr, Sternberg P Jr, et al. Subfoveal choroidal neovascularization in punctate inner choroidopathy. Surgical management and pathologic findings. Ophthalmology 1996;103:2061–9. 7. Gass JDM. Biomicroscopic and histopathologic considerations regarding the feasibility of surgical excision of subfoveal neovascular membranes. Am J Ophthalmol 1994;118: 285–98. 8. Green WR, Enger C. Age-related macular degeneration histopathologic studies. The 1992 Lorenz E. Zimmerman Lecture. Ophthalmology 1993;100:1519 –35. 9. Grossniklaus HE, Gass JDM. Clinicopathologic correlations of surgically excised type 1 and type 2 submacular choroidal neovascular membranes. Am J Ophthalmol 1998;126:59 – 69. 10. Melberg NS, Thomas MA, Burgess DB. The surgical removal of subfoveal choroidal neovascularization. Ingrowth site as a predictor of visual outcome. Retina 1996;16:190 –5. 11. Shiraga F, Ojima Y, Matsuo T, et al. Feeder vessel photocoagulation of subfoveal choroidal neovascularization secondary
Shiraga et al 䡠 Identification of Ingrowth Site of Idiopathic CNV by ICG
12. 13. 14.
15.
16.
to age-related macular degeneration. Ophthalmology 1998; 105:662–9. Destro M, Puliafito CA. Indocyanine green videoangiography of choroidal neovascularization. Ophthalmology 1989;96: 846 –53. Yannuzzi LA, Slakter JS, Sorenson JA, et al. Digital indocyanine green videoangiography and choroidal neovascularization. Retina 1992;12:191–223. Scheider A, Kaboth A, Neuhauser L. Detection of subretinal neovascular membranes with indocyanine green and an infrared scanning laser ophthalmoscope. Am J Ophthalmol 1992; 113:45–51. Chang TS, Freund KB, de la Cruz Z, et al. Clinicopathologic correlation of choroidal neovascularization demonstrated by indocyanine green angiography in a patient with retention of good vision for almost four years. Retina 1994;14:114 –24. Lee BL, Lim JI, Grossniklaus HE. Clinicopathologic features of indocyanine green angiography-imaged, surgically excised choroidal neovascular membranes. Retina 1996;16:64 –9.
17. Iida T, Hagimura N, Kishi S, Shimizu K. Indocyanine green angiographic features of idiopathic submacular choroidal neovascularization. Am J Ophthalmol 1998;126:70 – 6. 18. Lopez PF, Lambert HM, Grossniklaus HE, Sternberg P Jr. Well-defined subfoveal choroidal neovascular membranes in age-related macular degeneration. Ophthalmology 1993;100: 415–22. 19. Korte GE, Reppucci V, Henkind P. RPE destruction causes choriocapillary atrophy. Invest Ophthalmol Vis Sci 1984;25: 1135– 45. 20. Ho AC, Yannuzzi LA, Pisicano K, DeRosa J. The natural history of idiopathic subfoveal choroidal neovascularization. Ophthalmology 1995;102:782–9. 21. Argon laser photocoagulation for idiopathic neovascularization. Results of a randomized clinical trial. Arch Ophthalmol 1983;101:1358 – 61. 22. Krypton laser photocoagulation for idiopathic neovascular lesions. Results of a randomized clinical trial. Macular Photocoagulation Study Group. Arch Ophthalmol 1990;108: 832–7.
607