Subjects

Consecutive treatment-naïve patients with pathologic myopia (defined as a spherical equivalent of <−6.00 diopters and/or an axial length [AL] of >26 mm with chorioretinal abnormalities such as lacquer cracks, chorioretinal atrophy, and posterior staphyloma ) and exudative lesions who visited the Department of Ophthalmology and Visual Sciences at Kyoto University Graduate School of Medicine (Kyoto, Japan) between February 2015 and November 2015 were recruited for this study. All patients underwent comprehensive ophthalmologic examinations, including measurement of the best-corrected visual acuity (BCVA) using a decimal visual acuity chart (Landolt chart) and AL using an IOLMaster (Carl Zeiss Meditec, Inc, Dublin, California, USA). All BCVA data were converted to the logarithm of the minimal angle of resolution (logMAR) for statistical analyses. Indirect ophthalmoscopy, slit-lamp biomicroscopy, color fundus photography (TRC-NW8F; Topcon Corp, Tokyo, Japan), SDOCT (Spectralis HRA+OCT; Heidelberg Engineering, Heidelberg, Germany), FFA, indocyanine green angiography (ICGA) (HRA-2; Heidelberg Engineering), and OCTA (RTVue XR Avanti with AngioVue; Optovue, Inc, Fremont, California, USA) were also performed. OCTA images are obtained twice in approximately 3 seconds. Although OCTA images can be obtained without pupil dilation depending on the patient condition, we induced mydriasis before the procedure to obtain high-quality images. The inclusion criterion was as follows: presence of exudative lesions, including subretinal hemorrhage, serous retinal detachment, and retinal edema, associated with pathologic myopia diagnosed by retinal specialists using slit-lamp biomicroscopy, color fundus photography, and SDOCT. A total of 28 eyes with exudative lesions in 26 consecutive Japanese patients (22 women, 4 men) were eventually included. Among these, 23 eyes of 22 patients were diagnosed with myopic CNV (CNV group) and 5 eyes of the remaining 4 patients were diagnosed without CNV (simple hemorrhage group) by retinal specialists using biomicroscopy, color fundus photography, SDOCT, FFA, and ICGA.

Analysis of Choroidal Neovascularization With Fluorescein Angiography and Optical Coherence Tomography Angiography

All FFA images of the 28 eyes were independently evaluated by 2 investigators (M.M. and M.H.). CNV was defined by the presence of a hyperfluorescence on early-phase images that increased in size and intensity on late-phase images. The CNV area was measured on early-phase FFA images after correction for AL-related magnification by HRA-2 built-in software ( Figure 1 ).

(Top row) Representative fundus fluorescein angiography and optical coherence tomography angiography images for a 65-year-old female patient with an axial length of 28.63 mm and a small area of choroidal neovascularization and (Bottom row) a 58-year-old female patient with an axial length of 28.36 mm and choroidal neovascularization of a moderate size. (Top left and Bottom left) The left B-scan images are obtained using optical coherence tomography; (Top center and Bottom center) the central images are obtained using fundus fluorescein angiography; and (Top right and Bottom right) the right images are obtained using optical coherence tomography angiography. The measured choroidal neovascularization areas are (Top center) 0.04 mm ² , (Top right) 0.05 mm ² , (Bottom center) 0.44 mm ² , and (Bottom right) 0.34 mm ² . The structure of fine vessels is more clearly visible on (Top right and Bottom right) optical coherence tomography angiography images than on (Top center and Bottom center) fundus fluorescein angiography images. The optical coherence tomography B-scan image does not appear reliable for accurate detection of the size of choroidal neovascularization.

All OCTA images were obtained using RTVue XR Avanti with the AngioVue macular cube (3 × 3 mm) protocol. Each imaging cube comprised 304 clusters of repeated B-scans that contained 304 A-scans each. Retinal layer segmentation was automatically performed by the built-in software. Outer retinal level OCTA images were independently analyzed by 2 investigators to confirm the presence or absence of CNV. The CNV area was manually measured using ImageJ software (public domain software; National Institutes of Health, Bethesda, Maryland, USA; Figure 1 ). However, the RTVue XR Avanti system does not have a built-in correction system for AL-related magnification. Because the SDOCT scan area increases for eyes with a long AL, we had to consider adjusting the length of an image quadrant (304 pixels) on 1 side, particularly for eyes with high myopia in the present study. A square raster scan area measuring 10 × 10 degrees, which is equivalent to a square measuring 3 × 3 mm in the Gullstrand model eye, was used. The length on 1 side was corrected using AL, the flatter meridian, the steeper meridian, and the spherical equivalent refraction in the modified Littmann formula (Bennett formula).

Statistical Analyses

Data are presented as means ± standard deviations where applicable. All statistical analyses were performed using SPSS (version 21; IBM, Armonk, New York, USA). We used Student t tests or χ ² tests to compare datasets as appropriate and analyzed correlations using Pearson rank correlation coefficients. Intraclass correlation coefficients (ICCs) for the area of CNV determined by the 2 investigators were calculated to determine the reliability of both OCTA and FFA measurements in 10 patients. A P value of <.05 was considered statistically significant.