Subjects

We retrospectively reviewed the medical records of 297 eyes of 255 patients with treatment-naive myopic CNV who presented at Yonsei University Medical Center between January 2005 and December 2011. Informed consent was obtained from all eligible participants at the time of angiographic examination, regarding the purpose of the examinations and possible complications due to angiographic dyes. This retrospective study was approved by the Institutional Review Board of Yonsei University Medical Center (approval number 4-2012-0171) and adhered to the tenets of the Declaration of Helsinki.

Inclusion criteria were (1) patients initially diagnosed with myopic CNV; and (2) absence of concurrent ocular diseases in the enrolled eyes that might compromise FA and ICGA images in the study eye, such as retinal vein occlusion and diabetic retinopathy. Exclusion criteria were (1) any history of prior treatment for CNV, such as laser photocoagulation, submacular surgery, or photodynamic therapy; and (2) the absence of FA and ICGA data.

Definitions for this study were pathologic myopia as (1) a refractive error exceeding −6.5 diopters or an axial length exceeding 26.5 mm; and (2) fundus changes indicative of pathologic myopia, including lacquer cracks, atrophic patches, or chorioretinal atrophy. Myopic CNV was defined as (1) eyes exhibiting the characteristic features of pathologic myopia described above; (2) relatively localized, focal late-phase leakage on FA; and (3) absence of features suggesting AMD such as drusen. PCV was defined as the presence of abnormal branching vascular networks with terminating polypoidal lesions on ICGA. Recurrence of myopic CNV was defined as evidence of leakage from a previously closed CNV on FA images.

Type 1 occult CNV was defined as poorly defined forms of CNV on FA; fibrovascular pigment epithelial detachment (stippled hyperfluorescence within 1 to 2 minutes after fluorescein injection with persistent staining or leakage in the overlying subretinal space by 10 minutes), or late-phase leakage from an undetermined source (late speckled hyperfluorescence lacking well-demarcated borders, with pooling of the eye in the overlying subretinal space). The location of type 1 occult CNV was confined to the subRPE space by OCT images. Type 2 classic CNV was defined as well-demarcated hyperfluorescence with progressive pooling of dye leakage in the overlying subretina space in the late phase. The neovascular membrane of type 2 classic CNV proliferates above the RPE in the subneurosensory space as shown by OCT.

Treatments for myopic CNV included photodynamic therapy (PDT) and anti-VEGF therapy. During the earlier period of the study, patients were treated with PDT. With the introduction of anti-VEGF agents, both PDT and anti-VEGF therapy were used for myopic CNV, depending on the physician’s discretion. After late 2007, anti-VEGF therapy has been the mainstay of treatment for both treatment-naive and recurrent myopic CNV.

Examination

All patients underwent complete ophthalmologic evaluation, including color fundus photography, FA, and ICGA at baseline. OCT images were obtained using a time domain OCT device (Stratus Carl Zeiss Meditec, Dublin, CA, USA) for patients before 2009, and a spectral domain OCT device (SD OCT; Spectralis; Heidelberg Engineering, Heidelberg, Germany) from 2009 onward. FA and ICGA data were collected for all enrolled patients during a baseline evaluation using a Heidelberg retina angiograph (HRA-II; Heidelberg Engineering, Dossenheim, Germany). All ICGA images were reviewed by 2 independent examiners (H.M.K., H.J.K.) to determine the presence of polypoidal lesions, and good agreements were made. The greatest linear dimension of myopic CNV was measured and calculated by the early phase FA image at 1 minute, using embedded software programs in the HRA system.

Patient characteristics retrieved from their medical charts included age at initial diagnosis, gender, refractive errors, and axial length at the time of diagnosing myopic CNV. Refractive error in all patients was measured by an auto-refractometer without cycloplegia. The spherical equivalent of refractive error in diopters was used in the statistical analyses. The greatest linear dimension of myopic CNV was also retrieved from image analysis.

Statistical analysis was performed using SPSS 18.0 for Windows (SPSS, Chicago, IL, USA).

Myopic Choroidal Neovascularization at Initial Presentation

Among 297 eyes (255 patients) with myopic CNV, 232 eyes (192 patients, 75.3%) were female. Mean refractive errors and axial lengths were −11.95 ± 5.88D (range −39.00D to −5.25D) and 29.39 ± 2.02 mm (range 26.7 mm to 34.4 mm), respectively. Mean age at diagnosis was 47.32 ± 14.69 years, and 141 eyes (47.5%; 118 patients) were from patients aged over 50 years (mean age 60.48 ± 7.34 years).

Angiographic Imaging Studies at Initial Presentation

All myopic CNV eyes were type 2 CNV, and the mean greatest linear dimension of the lesions was 1794.25 ± 1157.14 μm. ICGA analysis showed no polypoidal component associated with CNV in any of the eyes. Representative eyes are shown in Figure 1 and Figure 2 .

Angiographic findings in a 41-year-old female patient with myopic choroidal neovascularization in the left eye. Refractive error was −9.5 diopters, and axial length was 27.58 mm in the affected eye. A well-demarcated hyperfluorescent lesion was detected on early-phase fluorescein angiography (top left). A dark rim surrounding a faint hyperfluorescent neovascular membrane, corresponding to choroidal neovascularization, was noted on early-phase indocyanine green angiography, but no polypoidal lesions were detected (top right). On late-phase fluorescein angiography, leakage from the neovascular membrane was noted (bottom left), and lacquer crack with neovascular membrane was noted on late-phase indocyanine green angiography (bottom right).

Fluorescein angiography and indocyanine green angiography of an eye with myopic choroidal neovascularization. This 34-year-old female patient had myopic choroidal neovascularization in the right eye, with a refractive error of −14.0 diopters and an axial length of 33.47 mm. Early-phase fluorescein angiography showed a well-demarcated neovascular membrane (top left), and the corresponding lesion was surrounded by a dark rim on early-phase indocyanine green angiography, without polypoidal components (top right). Leakage from the neovascular membrane on late-phase fluorescein angiography was noted (bottom left). On late-phase indocyanine green angiography, a neovascular membrane was noted with multiple lacquer cracks (bottom right).

Recurrences

Among 297 eyes, 243 eyes (81.8%) completed at least 1 year of follow-up after initial treatment. Mean follow-up duration was 58.67 ± 38.42 months. In 86 eyes (35.4%) we saw at least 1 recurrence, with the mean number of recurrences being 1.86 ± 0.67 times. The mean interval between first remission and first recurrence was 18.74 ± 17.33 months. ICGA imaging studies at recurrence showed no polypoidal lesions associated with recurred myopic CNV.