Myopia is a major cause of visual impairment in many

Transcription

1 Retina Association between Choroidal Morphology and Anti- Vascular Endothelial Growth Factor Treatment Outcome in Myopic Choroidal Neovascularization Seong Joon Ahn, 1 Se Joon Woo, 1 Ko Eun Kim, 2 and Kyu Hyung Park 1 PURPOSE. To investigate associations between outcome of antivascular endothelial growth factor (VEGF) therapy and choroidal morphology in eyes with myopic choroidal neovascularization (CNV). METHODS. Fifty-two eyes of 46 patients with myopic CNV received a single intravitreal anti-vegf injection, followed by as-needed injections. Baseline choroidal thickness was measured at the fovea and 1.5 and 3 mm nasal, temporal, superior, and inferior to the fovea using enhanced depth imaging optical coherence tomography. Measurements were compared between eyes with and without CNV resolution after a single injection and between those with and without CNV recurrence within 1 year of initial injection. Associations between treatment outcomes and morphologic or clinical factors were assessed using regression analyses. RESULTS. Patients received intravitreal injections during follow-up. Eyes with CNV resolution after a single anti- VEGF injection had a significantly thicker inferior choroid than those without resolution ( vs lm, P ¼ 0.002). The subfoveal choroid was thinner in eyes with recurring CNV than in those without recurrence ( vs lm, P ¼ 0.029). Associations were found between inferior choroidal thickness and CNV resolution (P ¼ 0.019) and between subfoveal choroidal thickness and 1-year recurrence rates (P ¼ 0.016). Adjusted odds ratios were 9.1 for CNV resolution with an inferior choroidal thickness >49 lm and 5.6 for recurrence within 1 year with a subfoveal choroidal thickness 47.5 lm. CONCLUSIONS. A thinner subfoveal/inferior choroid at baseline may indicate poor anatomic outcome after intravitreal anti- VEGF treatment in eyes with myopic CNV. (Invest Ophthalmol Vis Sci. 2013;54: ) DOI: /iovs From the 1 Department of Ophthalmology, Seoul National University College of Medicine, Seoul National University Bundang Hospital, Seongnam, Korea; and the 2 Department of Ophthalmology, Seoul National University College of Medicine, Seoul National University Hospital, Seoul, Korea. Supported in part by a grant from the Korea Health Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (Grant No. A111161), and the National Research Foundation of Korea Grant 2012R1A2A2A (SJW). Submitted for publication December 20, 2012; revised February 4 and February 19, 2013; accepted February 27, Disclosure: S.J. Ahn, None;S.J. Woo, None;K.E. Kim, None; K.H. Park, None Corresponding author: Se Joon Woo, Department of Ophthalmology, Seoul National University Bundang Hospital, #300, Gumidong, Bundang-gu, Seongnam, Gyeonggi-do Korea; sejoon1@snu.ac.kr. Myopia is a major cause of visual impairment in many countries. Pathologic myopia is characterized by progressive anteroposterior elongation of the sclera and is associated with diverse secondary ocular changes. Choroidal neovascularization (CNV) caused by pathologic myopia, that is, myopic CNV, is a serious, vision-threatening condition in these patients. Although the incidence of myopic CNV has not been extensively investigated, a hospital-based study reported an occurrence rate of 5% to 10% in highly myopic patients. 1 Among secondary causes of CNV, myopia is the most common, accounting for 62% of all CNV cases in patients less than 50 years of age. 1 High-resolution cross-sectional images obtained with spectral-domain optical coherence tomography (SD-OCT) can be used to detect morphologic changes in the choroid and sclera in myopic eyes. 2 4 Ikuno et al. 4 identified several morphologic abnormalities of the choroid and suggested that some were risk factors for myopic CNV. Enhanced depth imaging SD-OCT (EDI- OCT) shows even more details of choroidal changes in myopic eyes. Fujiwara et al. 2 showed a thinned choroid in myopic eyes and determined the association between age-related choroidal thinning and the degree of myopia. Choroidal thickness measurements obtained using high-penetration OCT (HP- OCT) show good agreement with those obtained using EDI- OCT, 5 and Maruko et al. 6 showed unique choroidal and scleral characteristics in eyes with pathologic myopia using HP-OCT. The findings of these studies suggest that pronounced mechanical stretching, and the associated choroidal thinning, play a role in the development of myopic CNV. 4,6 Although studies have presented evidence of a pathogenic association between choroidal morphologic characteristics and myopic CNV, 4,6 little is known about the effect of these characteristics on the outcome of intravitreal anti-vascular endothelial growth factor (VEGF) therapy, the most common treatment for myopic CNV Here, we evaluate associations of baseline choroidal morphologic parameters with treatment outcome, CNV resolution, and CNV recurrence within 1 year of intravitreal anti-vegf therapy in eyes with myopic CNV. Because anatomic outcomes may be associated with visual outcome, associations between choroidal morphologic factors and final visual acuity were also examined. METHODS Subjects This retrospective study included 62 eyes from 56 consecutive patients who visited Seoul National University Bundang Hospital between January 2009 and June 2012 with high myopia (axial length >26.5 mm 4,6 9,12,13 and/or refractive error < dipoters [D] 4,11 ) and newly developed myopic CNV. All patients had received intravitreal anti-vegf injections and had undergone follow-up examinations for at least 3 months. Approval to conduct this study was obtained from the Investigative Ophthalmology & Visual Science, March 2013, Vol. 54, No. 3 Copyright 2013 The Association for Research in Vision and Ophthalmology, Inc. 2115

2 2116 Ahn et al. IOVS, March 2013, Vol. 54, No. 3 FIGURE 1. A typical enhanced depth imaging optical coherence tomography (EDI-OCT) image showing morphologic parameters measured in this study. White double arrows indicate choroidal thickness. In addition to the subfovea, choroidal thickness was measured 1.5 and 3 mm away from the fovea in each anatomic direction. Institutional Review Board (IRB) of Seoul National University Bundang Hospital, and the study adhered to the tenets of the Declaration of Helsinki. Myopic CNV was defined as the presence of dye leakage associated with myopic changes on a fluorescein angiogram (FA). Thorough fundus examination and SD-OCT images (Spectralis; Heidelberg Engineering, Heidelberg, Germany) confirmed that no other macular disease was present, particularly age-related macular degeneration, foveoschisis, macular hole, or epiretinal membrane. Additionally, patients who were treated with photodynamic therapy (PDT) or other intravitreal injections were excluded from analyses (n ¼ 3). Patients older than 75 years (n ¼ 2) were also excluded because the etiology of CNV in these patients may not have been exclusively myopic. We also excluded patients in whom the retinal pigment epithelium (RPE), choroid, and sclera were indistinguishable (n ¼ 3). Thus, 54 eyes from 48 patients were ultimately included in our analyses. Examinations Before treatment, a complete ophthalmic evaluation was performed for each patient, including best-corrected visual acuity (BCVA) assessment, slit-lamp biomicroscopy, intraocular pressure (IOP) measurement, fundus photography, FA, indocyanine green angiography (ICGA), and SD-OCT. Trained examiners, masked to patients information, measured BCVA in the examination room using a Snellen chart at 6 m. Fundus photographs, FA, and ICGA were used to evaluate CNV location and to grade myopic degeneration (scale: M0 M5) and lacquer cracks (scale: 0 2), according to the methods described by Avila et al. 14 and Ikuno et al., 15 respectively. Late-phase ICGA images, obtained using the Heidelberg Retina Angiograph 2 (HRA2; Heidelberg Engineering), were used for evaluation of lacquer cracks. Full-thickness choroidal images were obtained using EDI-OCT with eye-tracking and image-averaging systems, as described by Spaide et al. 3 The OCT images obtained before anti-vegf therapy were used for baseline analyses. The device was operated by a single experienced technician, who positioned the OCT camera close enough to the eye to obtain an inverted image. The eye-tracking and image-averaging capabilities of the device enabled better visualization of the choroid and increased the signal-to-noise ratio. Choroidal thickness was measured manually with calipers as the distance from the outer border of the RPE to the inner surface of the sclera (Fig. 1). These measurements were obtained at the fovea and 1.5 and 3 mm superior, inferior, nasal, and temporal to the fovea. All measurements were obtained on horizontal and vertical OCT scans that passed through the fovea. 6 OCT interpretations and measurements were performed by two independent, experienced investigators (SJA, KEK) who were masked to patient information. The average of the investigators measurements was calculated and used in analyses. When needed, a retina specialist (SJW) was consulted. If a convex elevation of the macula was observed on OCT images, a dome-shaped macula was diagnosed. 16 During the follow-up period, patients were seen monthly. At each visit, BCVA was measured and OCT was performed in the treated eye. In patients with symptoms or signs of CNV aggravation, FA was also performed. Treatment outcome, assessed as resolved or not resolved, was evaluated 1 month after the initial anti-vegf injection using OCT and FA. Resolution was defined as absence of intra-/subretinal fluid, as identified on OCT images, and no fluorescein leakage. Recurrence of CNV was defined, also on an anatomic basis, as the recurrence of intra-/ subretinal fluid, as detected on OCT images, and fluorescein leakage. Treatment Patients were treated with a single 0.05 ml intravitreal injection of an anti-vegf agent (1.25 mg bevacizumab [Avastin; Genentech, San Francisco, CA] or 0.50 mg ranibizumab [Lucentis; Novartis, Basel, Switzerland]) at baseline. Patients were re-treated with the drug that was used in the initial injection on an as-needed basis. Retreatment was administered for recurrent or aggravated intra-/subretinal fluid on OCT. All intravitreal injections were administered under topical anesthesia using a 30-gauge needle inserted 3.5 to 4.0 mm posterior to the limbus. Injections were administered in an outpatient setting using strict aseptic techniques. Statistical Analysis Descriptive statistics were obtained for data pertaining to demographics, axial length, CNV location, lacquer crack grade, BCVA at baseline, presence/absence of posterior staphyloma, and presence/absence of a dome-shaped macula. For descriptive statistics of choroidal thickness data, a box plot was used to display the distribution. Relationships between various choroidal morphologic parameters were investigated using linear regression analyses. The intraclass correlation coefficient (ICC) was used to examine interobserver agreement in choroidal thickness measurements. Pearson s correlation coefficients were also calculated. Clinical and morphologic factors associated with anatomic outcome were evaluated using Student s t-test or the Mann-Whitney U test, depending on normality, as determined by results of the Shapiro-Wilk test. These methods were used to compare continuous variables between eyes with and without CNV resolution after a single intravitreal anti-vegf injection. The same tests were also used to compare clinical and morphologic parameters between eyes with and without recurrence of intra-/subretinal fluid within 12 months of the initial injection. Predictive factors for CNV resolution after a single injection and for 1-year recurrence were identified using logistic regressions. Predictability of clinical and morphologic parameters for anatomic outcome was evaluated by receiver operating characteristic (ROC) curve analysis, and sensitivity and specificity were determined using cutoff values. Clinical factors (initial BCVA, age, axial length, and central

3 IOVS, March 2013, Vol. 54, No. 3 Choroid and Anti-VEGF Outcome in Myopic CNV 2117 TABLE 1. Demographics, Clinical Characteristics, and Treatment Details of Included Patients Factors Values Number of eyes 54 Age, y (range, 37 74) Sex, female (%) 44 (81.5) Follow-up period, mo (range, 3 42) Spherical refractive error, D Axial length, mm Best-corrected visual acuity at baseline, logmar Central macular thickness, lm Location of CNV, foveal:extrafoveal (%) 42:12 (77.8:22.2) Lacquer crack grade, 0:1:2 (%) 5:31:18 (9.3:57.4:33.3) Myopic degeneration grade, 0:1:2:3:4:5 (%) 0:1:12:23:18:0 (0:1.9:22.2:42.6:33.3:0) Posterior staphyloma (%) 51 (94.4) Dome-shaped macula (%) 9 (16.7) Materials used for injection, bevacizumab:ranibizumab (%) 41:13 (75.9:24.1) Number of injections during follow-up periods (range, 1 5) macular thickness [CMT]) and choroidal morphologic parameters were used to calculate the area under the ROC curve (AUC) for resolution or recurrence within 1 year. Cutoff values for stratifying patients with myopic CNV into groups with thicker and thinner choroid were determined to maximize sensitivity for the prediction of resolution and 1-year recurrence without significant loss of specificity (around 80%). The cutoff values were used to calculate the odds ratios (ORs) for CNV resolution and CNV recurrence. Associations between clinical and morphologic parameters and visual outcome (final BCVA and change in BCVA from baseline) were evaluated using linear regression analyses. Possible predictive factors of BCVA were assessed using a multiple regression analysis. Continuous values are expressed as mean 6 standard deviation. Statistical significance was defined as P < All statistical tests were performed using SPSS software (version 18.0; SPSS, Inc., Chicago, IL). RESULTS Demographic and Clinical Characteristics Demographic data from 54 eyes (48 patients) are summarized in Table 1. The mean patient age was years, and the mean follow-up period was months (range, 3 42 months). Mean spherical refractive error was D, and mean axial length was mm. Lacquer cracks were present in 48 of 54 (88.9%) eyes with myopic CNV. Foveal and extrafoveal CNVs were present in 42 (77.8%) and 12 (22.2%) eyes, respectively. The location of CNV was within the lacquer crack or adjacent to it in 32 (59.3%) and 11 (20.4%) eyes, respectively. The CNV appeared independent of the lacquer crack in 11 (20.4%) eyes. Choroidal Thickness Figure 2 shows the descriptive statistics, represented by box plots, of baseline choroidal thickness in each region examined. The nasal choroid was the thinnest area, whereas the superior choroid was the thickest area. The subfoveal choroid was thinner than the superior, inferior, and temporal choroid. Subfoveal choroidal thickness was inversely correlated with axial length (r ¼ 0.285, P ¼ 0.079), but this correlation was only marginally significant. However, the thickness was strongly correlated with choroidal thickness in other regions FIGURE 2. Box plots showing choroidal thicknesses measured in this study. The nasal choroid was the thinnest area, whereas the superior choroid was the thickest area. The subfoveal choroid was thinner than the superior, inferior, and temporal choroid.

4 2118 Ahn et al. IOVS, March 2013, Vol. 54, No. 3 TABLE 2. Comparison of Clinical and Morphologic Factors between Patients with and without Myopic CNV Resolution Treatment Response to Anti-VEGF Without Resolution, n ¼ 18 With Resolution, n ¼ 36 P Age * Baseline BCVA * Baseline CMT * Spherical equivalent, D * Axial length, mm * Materials used for initial treatment, bevacizumab (%) 12 (66.7) 29 (80.6) CNV location, foveal (%) 15 (83.3) 27 (75.0) Lacquer crack grade, 0:1:2 2:10:6 3:21: Myopic degeneration grade, 1:2:3:4:5 1:2:10:5:0 0:10:13:13: Posterior staphyloma (%) 17 (94.4) 34 (94.4) 1.0 Dome-shaped macula (%) 4 (22.2) 5 (13.9) Choroidal thickness, lm Subfovea * 3 mm superior * 3 mm inferior * 3 mm nasal mm temporal mm superior * 1.5 mm inferior * 1.5 mm nasal * 1.5 mm temporal * Statistical analyses were performed using Student s t-test or the Mann-Whitney U test for continuous variables depending on normality, as determined by results of the Shapiro-Wilk test, and v 2 or Fisher s exact test for dichotomous or interval variables. * Student s t-test. Mann-Whitney U test. (P < for 1.5 and 3 mm superior, inferior, nasal, and temporal areas). There was good interobserver agreement in choroidal thickness measurements, with ICCs ranging between (nasal 3 mm: 95% CI, ) and (subfoveal: 95% CI, ). The ICC for inferior 3 mm choroidal thickness measurements was (95% CI, ). Pearson s correlation coefficient ranged between and (P < 0.001). Anatomic Outcomes and Associated Factors Intravitreal injections of bevacizumab and ranibizumab were administered in 41 (75.9%) and 13 (24.1%) eyes, respectively. The mean number of injections was (range, 1 4) during the first year and (range, 1 5) over the entire follow-up period. After a single anti-vegf agent injection, 36 (66.7%) eyes showed resolution of myopic CNV, with no residual intra- or subretinal fluid. Table 2 and Figure 3 show the comparison of clinical and baseline morphologic characteristics between patients who showed resolution of myopic CNV and those who did not. Three millimeters inferior to the fovea, the choroid was significantly thicker in eyes with resolution ( lm) than in eyes without resolution ( lm; P ¼ 0.002, Student s t-test). Additionally, baseline BCVA was significantly better in eyes with CNV resolution ( logmar) than in eyes without resolution ( logmar, P ¼ 0.044). Logistic regression analysis suggested that baseline choroidal thickness 3 mm inferior to the fovea was significantly associated with CNV resolution (OR ¼ per 1 lm increment; 95% CI, ; P ¼ 0.019). This result indicates that eyes with a thicker inferior choroid were more likely to have CNV resolution after a single anti-vegf treatment. Among the 39 eyes that underwent follow-up examinations for more than 1 year, 19 (48.7%) needed retreatment within the year because of intra-/subretinal fluid recurrence. Table 3 and Figure 4 show comparisons of clinical and baseline morphologic characteristics between patients with and without recurrence within 1 year. Subfoveal choroidal thickness was significantly different between eyes with ( lm) and without ( lm) recurrence (P ¼ 0.029, Student s t- test). Additionally, foveal CNV was more frequently noted in eyes with recurrence within 1 year (89.5% vs. 60%, P ¼ 0.039). Logistic regression analyses revealed that baseline subfoveal choroidal thickness was the only predictive factor for CNV recurrence within 1 year (OR ¼ per 1 lm increment; 95% CI, ; P ¼ 0.016). This result indicates that recurrence was less likely to occur in eyes with a thicker subfoveal choroid. Predictability of Clinical and Choroidal Parameters for Anatomic Outcome Inferior choroidal thickness was the strongest predictive morphologic parameter for CNV resolution after a single anti- VEGF injection (AUC, 0.78 and 0.66 at 3 and 1.5 mm inferior to the fovea, respectively). Clinical parameters such as initial BCVA, age, axial length, spherical equivalent, and CMT showed smaller AUC values (0.66, 0.54, 0.47, 0.51, and 0.50, respectively). However, for predicting recurrence within 1 year, subfoveal choroidal thickness was the most useful parameter (AUC, 0.79). After separation of eyes into two groups based on inferior choroidal thickness 3 mm from the fovea (cutoff value, 49 lm), analyses revealed a 77.8% sensitivity and 83.3% specificity for complete CNV resolution after a single anti-vegf injection. A cutoff value of 47.5 lm for subfoveal choroidal thickness resulted in a 66.7% sensitivity and 80.0% specificity for predicting recurrence of CNV within 1 year.

5 IOVS, March 2013, Vol. 54, No. 3 Choroid and Anti-VEGF Outcome in Myopic CNV 2119 FIGURE 3. Comparison of choroidal thickness with and without resolution of myopic choroidal neovascularization after a single intravitreal injection of an anti-vascular endothelial growth factor agent. Choroidal thickness was measured 3 and 1.5 mm from the fovea in each anatomic direction. Error bars indicate the upper bound of 95% confidence intervals. The asterisk indicates a statistically significant difference (P ¼ by Student s t-test). TABLE 3. Injection Comparison of Clinical and Morphologic Factors between Eyes with and without Recurrence within 1 Year after the Initial Anti-VEGF Clinical/Morphologic Factors Nonrecurrence, n ¼ 20 Recurrence, n ¼ 19 P Age * Baseline BCVA * Baseline CMT * Spherical equivalent, D * Axial length, mm * Materials used for treatment, bevacizumab (%) 14 (70.0) 15 (78.9) CNV location, foveal (%) 12 (60.0) 17 (89.5) Lacquer crack grade, 0:1:2 2:12:6 2:10: Myopic degeneration grade, 1:2:3:4:5 0:5:8:7:0 1:3:8:7: Posterior staphyloma (%) 19 (95) 18 (94.7) Dome-shaped macula (%) 3 (15.0) 3 (15.8) Choroidal thickness, lm Subfovea * 3 mm superior * 3 mm inferior * 3 mm nasal * 3 mm temporal mm superior mm inferior * 1.5 mm nasal * 1.5 mm temporal * Statistical analyses were performed using Student s t-test or the Mann-Whitney U test for continuous variables depending on normality, as determined by results of the Shapiro-Wilk test, and v 2 or Fisher s exact test for dichotomous or interval variables. * Student s t-test. Mann-Whitney U test.

6 2120 Ahn et al. IOVS, March 2013, Vol. 54, No. 3 FIGURE 4. Comparison of choroidal thickness between eyes with and without recurrence of choroidal neovascularization within 1 year after a single intravitreal injection of an anti-vascular endothelial growth factor agent. Choroidal thickness was measured 3 and 1.5 mm from the fovea in each anatomic direction. Error bars indicate the upper bound of 95% confidence intervals. The asterisk indicates a statistically significance difference (P ¼ by Student s t-test). In eyes that had an inferior choroidal thickness (3 mm from the fovea) >49 lm, the OR (adjusted for age, sex, and potential confounders including subfoveal choroidal thickness) for complete CNV resolution after the initial injection was 9.1 (95% CI, , P ¼ 0.003). In those with subfoveal choroidal thickness 47.5 lm, the adjusted OR for recurrence within 1 year was 5.6 (95% CI, , P ¼ 0.017). Visual Outcomes and Associated Factors Intravitreal injections of an anti-vegf agent resulted in a significant visual improvement from at baseline to logmar (P ¼ 0.015, paired t-test). Table 4 presents the association between choroidal thickness in several regions and visual outcome. Baseline BCVA and CMT were significantly associated with final BCVA, but no choroidal thickness parameters were associated with final BCVA or change in BCVA from baseline. Among clinical factors, the only predictive factor of final BCVA was baseline BCVA, as identified by multiple regression analysis (r ¼ 0.71, P < 0.001). DISCUSSION Our study investigated the association between outcome of intravitreal anti-vegf injections and choroidal morphology. By measuring baseline choroidal thickness at multiple posterior pole locations using EDI-OCT images, we found that choroidal thinning 3 mm inferior to the fovea was associated with incomplete resolution of myopic CNV after a single anti-vegf injection. Additionally, subfoveal choroidal thinning was associated with 1-year recurrence of myopic CNV. According to ROC analyses, the subfoveal and inferior choroidal thickness likely predicts treatment outcome, resolution, and 1-year recurrence of CNV. Use of thickness cutoff values facilitates prediction of treatment outcomes after intravitreal anti-vegf therapy in clinical practice. We showed highly significant OR values (9.1 for resolution and 5.6 for recurrence within 1 year) when eyes were divided into two groups according to inferior and subfoveal choroidal thickness. In addition to axial length elongation in pathologic myopia, morphologic parameters of the choroid are believed to be associated with the development of myopic CNV. For example, Ikuno et al. 4 compared choroidal morphology of eyes with myopic CNV to that of healthy fellow eyes and showed TABLE 4. The Association between Visual Outcome after Intravitreal Anti-VEGF Injection and Clinical and Morphologic Parameters at Baseline Parameter Final BCVA BCVA Change b P b P Axial length Age Baseline BCVA < Baseline CMT Choroidal thickness Subfovea mm superior mm inferior mm nasal mm temporal mm superior mm inferior mm nasal mm temporal b, correlation coefficient.

7 IOVS, March 2013, Vol. 54, No. 3 Choroid and Anti-VEGF Outcome in Myopic CNV 2121 significant thinning of the subfoveal and inferior choroid in the eyes with myopic CNV. They also found that inferior choroidal thickness was an ocular risk factor for the development of myopic CNV. 4 It remains unknown why choroidal thinning is associated with development of myopic CNV. McLeod et al. 17 showed choriocapillaris dropout adjacent to active CNV in agerelated macular degeneration. They also showed an intimate relationship between RPE atrophy and choriocapillaris degeneration. Moriyama et al. 13 reported that 5 of 59 (8.5%) highly myopic eyes had narrowing of the large choroidal vein and that 2 of 59 (3.4%) eyes had complete loss of the large choroidal vein, 1 of which developed CNV. In eyes with myopic CNV, we hypothesize that RPE/Bruch s membrane defects associated with choriocapillaris degeneration and choroidal thinning lead to breakdown of the blood retinal barrier and subsequent CNV. Further studies on RPE changes in locations with choroidal thinning and their association with the development of CNV may be helpful for elucidating the pathogenesis of myopic CNV. The association between choroidal thinning and development of myopic CNV leads to the hypothesis that outcome of intravitreal anti-vegf injection therapy is associated with choroidal morphologic parameters. To further examine this hypothesis, we examined baseline choroidal thickness at the posterior pole in several regions for an association with outcome of anti-vegf therapy. We found that inferior choroidal thinning, a known risk factor for myopic CNV, 4 was also associated with incomplete resolution of CNV after anti-vegf therapy. We also showed that subfoveal choroidal thinning in eyes with myopic CNV showed significant association with 1- year recurrence of CNV. Previous studies 4 indicated that in eyes with myopic CNV, the height of the posterior staphyloma edge from the most posterior part of the eye was lower in the inferior sclera than in the other quadrants (nasal, superior, and temporal). This indicates that myopic elongation occurs along the visual axis of the eye and that the eye is most elongated subfoveally and in the inferior direction. Thus, subfoveal and inferior choroidal thinning may be manifestations of mechanical stretching caused by axial length elongation. However, because the shape of the posterior staphyloma can vary greatly, 18 further studies are needed to confirm the association between subfoveal/inferior choroidal thinning and treatment outcome. Calvo-Gonzalez et al. 10 investigated which clinical factors may be predictive of visual outcome after intravitreal ranibizumab injection in eyes with myopic CNV. They found that baseline BCVA and myopic CNV location were predictive of visual outcome. 10 Our study also showed that baseline BCVA was associated with final BCVA. In both studies, patients with better baseline BCVA were more likely to have better posttreatment visual outcomes. However, no choroidal morphologic factors were significantly associated with visual outcome or visual gain. Although choroidal thickness parameters may be valuable in predicting anatomic outcome of anti- VEGF treatment in eyes with myopic CNV, they had limited predictive value for visual outcome. Therefore, both inferior and subfoveal choroidal thickness measurements, from EDI- OCT images, and the usual clinical assessments, including BCVA measurement, may be valuable in predicting both visual and anatomic outcomes after intravitreal anti-vegf therapy in eyes with myopic CNV. Although the dosing regimen of intravitreal anti-vegf and other drugs has varied between studies, our patients were treated with a single anti-vegf injection, followed by asneeded retreatment during follow-up, as reported by Silva et al. 11 and Mones et al. 19 The visual gains in our study (7.8 Early Treatment Diabetic Retinopathy Scale [ETDRS] letters, as converted from a gain of logmar 20 ) were similar to those in these two previous reports ( and ETDRS letters). Our mean number of injections during 1 year (1.5) was also comparable to that reported by Mones et al. (1.52). 19 The appropriate anti-vegf dosing regimen remains controversial; and although our study did not compare the visual and anatomic outcome between different dosing schedules, our data support the use of a single anti-vegf treatment, followed by as-needed injections. Using this regimen, we observed a high percentage of complete CNV resolution after only a single anti-vegf treatment while reducing the number of intravitreal injections. This minimizes patient discomfort and decreases the risk of injection-related complications, such as endophthalmitis and retinal detachment. Additionally, our finding that choroidal thinning 3 mm inferior to the fovea is associated with incomplete resolution of CNV may aid clinicians in determining the initial anti-vegf loading dose (i.e., single or multiple injections). For eyes with subfoveal choroidal thinning, our study suggests that careful and regular follow-up is necessary to detect CNV recurrence within 1 year of initial treatment. Our study has several limitations. First, not all of our cases were treated with the same anti-vegf drug. Although recent studies 12 have suggested that efficacy of bevacizumab and ranibizumab is similar in terms of visual gain, the use of two different drugs may add a source of bias. We found that there were no significant differences in visual or anatomic outcomes between eyes treated with bevacizumab and those treated with ranibizumab (see Supplementary Material and Supplementary Table S1, /-/DCSupplemental). As there were no significant differences in baseline choroidal thickness at any point measured between bevacizumab- and ranibizumab-treated eyes, the use of different anti-vegf drugs likely did not confound our results. Second, because a dome-shaped macula generally has a thicker subfoveal choroid, 16 the nine eyes with a dome-shaped macula in our study may have influenced the association between subfoveal choroidal thickness and treatment outcome. In addition, the location and size of myopic CNV, in relation to our measuring points, may have affected our measurements; but these factors could not be controlled in our study. The interval between CNV development and examination may have influenced our study outcomes. We included patients who had at least 3 months of follow-up, but five patients were followed for less than 6 months. This short follow-up duration may have been insufficient to allow adequate determination of anatomic outcome in these patients. Additionally, we measured Snellen visual acuity, which has well-documented limitations, including inconsistent progression in letter size and unequal legibility of letters. Furthermore, this was a retrospective study with a small sample size, and larger, prospective studies are needed. Moreover, choroidal thickness can be affected by age, spherical equivalent, refractive error, axial length, and diurnal variation Factors affecting choroidal thickness must also be considered in interpretation of our choroidal thickness data. In particular, diurnal variation may have affected our results. Using Spectralis OCT, Tan et al. 24 showed that subjects with a thin choroid (300 lm) had a smaller diurnal variation in subfoveal thickness (11.8 lm) than the average in normal, healthy subjects (31.6 lm), indicating that the amplitude of diurnal variation may depend on choroidal thickness. Because our patients with myopic CNV had a very thin choroid (mean subfoveal thickness ¼ 46.7 lm), we expected diurnal variation in our patients to be less than 11.8 lm; and the difference in subfoveal choroidal thickness between recurrent and nonrecurrent cases was 17 lm, greater than the expected diurnal variation. Additionally, OCT images were acquired in each group of patients, on average, at almost same time of day (12:18 PM with resolution versus 11:56 AM without resolution;

8 2122 Ahn et al. IOVS, March 2013, Vol. 54, No. 3 12:06 PM with 1-year recurrence versus 11:57 AM without 1- year recurrence). As choroidal thickness peaks in the morning and progressively decreases during the day, 24 patients who have CNV resolution are expected to have a thinner choroid due to diurnal variation; however, our results showed the opposite. Patients with recurrent CNV are also expected to have a thinner subfoveal choroid due to diurnal variation, but the slight time difference in OCT acquisition (9 minutes) between patients with and without recurrence likely did not affect our results meaningfully. Finally, because choroidal thickness measurements can vary with different OCT devices, 25 our choroidal thickness cutoff values may be applicable only to Spectralis OCT measurements. In conclusion, inferior and subfoveal choroidal thinning at baseline were associated with incomplete resolution and recurrence after intravitreal anti-vegf therapy in eyes with myopic CNV. This association also supports the hypothesis of a pathogenic association between choroidal thinning and myopic CNV, as previously suggested by others. 4,6 References 1. Cohen SY, Laroche A, Leguen Y, Soubrane G, Coscas GJ. Etiology of choroidal neovascularization in young patients. Ophthalmology. 1996;103: Fujiwara T, Imamura Y, Margolis R, Slakter JS, Spaide RF. Enhanced depth imaging optical coherence tomography of the choroid in highly myopic eyes. Am J Ophthalmol. 2009;148: Spaide RF, Koizumi H, Pozzoni MC. Enhanced depth imaging spectral-domain optical coherence tomography. Am J Ophthalmol. 2008;146: Ikuno Y, Jo Y, Hamasaki T, Tano Y. Ocular risk factors for choroidal neovascularization in pathologic myopia. Invest Ophthalmol Vis Sci. 2010;51: Ikuno Y, Maruko I, Yasuno Y, et al. Reproducibility of retinal and choroidal thickness measurements in enhanced depth imaging and high-penetration optical coherence tomography. Invest Ophthalmol Vis Sci. 2011;52: Maruko I, Iida T, Sugano Y, Oyamada H, Akiba M, Sekiryu T. Morphologic analysis in pathologic myopia using highpenetration optical coherence tomography. Invest Ophthalmol Vis Sci. 2012;53: Parodi MB, Iacono P, Papayannis A, Sheth S, Bandello F. Laser photocoagulation, photodynamic therapy, and intravitreal bevacizumab for the treatment of juxtafoveal choroidal neovascularization secondary to pathologic myopia. Arch Ophthalmol. 2010;128: Gharbiya M, Allievi F, Mazzeo L, Gabrieli CB. Intravitreal bevacizumab treatment for choroidal neovascularization in pathologic myopia: 12-month results. Am J Ophthalmol. 2009; 147:84 93, e Ikuno Y, Sayanagi K, Soga K, et al. Intravitreal bevacizumab for choroidal neovascularization attributable to pathological myopia: one-year results. Am J Ophthalmol. 2009;147:94 100, e Calvo-Gonzalez C, Reche-Frutos J, Donate J, Fernandez-Perez C, Garcia-Feijoo J. Intravitreal ranibizumab for myopic choroidal neovascularization: factors predictive of visual outcome and need for retreatment. Am J Ophthalmol. 2011; 151: Silva RM, Ruiz-Moreno JM, Rosa P, et al. Intravitreal ranibizumab for myopic choroidal neovascularization: 12-month results. Retina. 2010;30: Iacono P, Parodi MB, Papayannis A, et al. Intravitreal ranibizumab versus bevacizumab for treatment of myopic choroidal neovascularization. Retina. 2012;32: Moriyama M, Ohno-Matsui K, Futagami S, et al. Morphology and long-term changes of choroidal vascular structure in highly myopic eyes with and without posterior staphyloma. Ophthalmology. 2007;114: Avila MP, Weiter JJ, Jalkh AE, Trempe CL, Pruett RC, Schepens CL. Natural history of choroidal neovascularization in degenerative myopia. Ophthalmology. 1984;91: Ikuno Y, Sayanagi K, Soga K, et al. Lacquer crack formation and choroidal neovascularization in pathologic myopia. Retina. 2008;28: Imamura Y, Iida T, Maruko I, Zweifel SA, Spaide RF. Enhanced depth imaging optical coherence tomography of the sclera in dome-shaped macula. Am J Ophthalmol. 2011;151: McLeod DS, Grebe R, Bhutto I, Merges C, Baba T, Lutty GA. Relationship between RPE and choriocapillaris in age-related macular degeneration. Invest Ophthalmol Vis Sci. 2009;50: Curtin BJ. The posterior staphyloma of pathologic myopia. Trans Am Ophthalmol Soc. 1977;75: Mones JM, Amselem L, Serrano A, Garcia M, Hijano M. Intravitreal ranibizumab for choroidal neovascularization secondary to pathologic myopia: 12-month results. Eye (Lond). 2009;23: , quiz Gregori NZ, Feuer W, Rosenfeld PJ. Novel method for analyzing snellen visual acuity measurements. Retina. 2010; 30: Brown JS, Flitcroft DI, Ying GS, et al. In vivo human choroidal thickness measurements: evidence for diurnal fluctuations. Invest Ophthalmol Vis Sci. 2009;50: Ikuno Y, Kawaguchi K, Nouchi T, Yasuno Y. Choroidal thickness in healthy Japanese subjects. Invest Ophthalmol Vis Sci. 2010;51: Margolis R, Spaide RF. A pilot study of enhanced depth imaging optical coherence tomography of the choroid in normal eyes. Am J Ophthalmol. 2009;147: Tan CS, Ouyang Y, Ruiz H, Sadda SR. Diurnal variation of choroidal thickness in normal, healthy subjects measured by spectral domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2012;53: Yamashita T, Shirasawa M, Arimura N, Terasaki H, Sakamoto T. Repeatability and reproducibility of subfoveal choroidal thickness in normal eyes of Japanese using different SD-OCT devices. Invest Ophthalmol Vis Sci. 2012;53: