In this retrospective case-control study, we reviewed patients diagnosed with nAMD with type 1 choroidal neovascularization (CNV) at Chungbuk National University Hospital from January 2017 to June 2020. This study was approved by the Institutional Review Board of the Chungbuk National University Hospital, and it adhered to the tenets of the Declaration of Helsinki. The need for informed consent from the participants was waived by Institutional Review Board of the Chungbuk National University Hospital due to the retrospective nature of this study. For the diagnosis, all participants underwent an ophthalmic examination, including a best-corrected visual acuity (BCVA) examination, slit-lamp examination, fundus photography, OCT, fluorescein angiography (FA), and indocyanine angiography (ICGA).
The inclusion criteria were as follows: (1) patients over 50 years of age with clinical features of nAMD; (2) patients diagnosed with type 1 CNV associated with subretinal fluid (SRF) and fibrovascular pigment epithelial detachment (PED) as per OCT;11 (3) patients who completed the initial 3-month loading dose treatment with intravitreal anti-VEGF agent—aflibercept (Eylea, Regeneron, Armonk, NY, USA); (4) patients who had not received any previous therapy, including laser, photodynamic therapy, or anti-VEGF treatment; and (5) patients who had been followed-up for 1-year. The exclusion criteria were as follows: (1) patients diagnosed with type 2 or 3 CNV; (2) patients with secondary CNV from another cause, including pathologic myopia or chorioretinitis; (3) SRF from pachychoroid neovasculopathy or central serous chorioretinopathy; (4) patients with complete retinal layer thinning, which meets the criteria for complete retinal pigment epithelium (RPE) and outer retinal atrophy due to extensive retinal destruction12; (5) patients with comorbidities such as other retinal or choroidal vascular disorder(s), including diabetic retinopathy, retinal vein occlusion, etc.; and (6) patients with poor quality of images, which precluded measurement of retinal layer thicknesses.
The CNV types were determined by the following criteria: type 1 CNVs were within the sub-RPE space, typically corresponding to angiographically occult CNV; type 2 CNVs were within the subretinal space, typically corresponding to angiographically classic CNV; and type 3 CNVs displayed intraretinal angiomatous proliferation11. All patients underwent multimodal imaging such as OCT, FA, and ICGA. If OCT was insufficient to determine the CNV type, then FA and ICGA were used as the reference. Intense hyperfluorescence that appeared early and showed progressive leakage was considered classic CNV and excluded. Staining/leakage with stippled hyperfluorescence and poorly demarcated areas of leakage in the late phase of FA was considered occult CNV and included. We only included type 1 CNVs in this study because ONL thickness measurement was impossible due to the disruption and destruction of the retinal laminar structure in the other CNV types. CNVs containing type 1 and type 2 lesions in a similar amount were classified as “type 2” and excluded. However, CNVs with predominantly type 1 characteristics with stippled leakage and staining pattern on FA without any retinal infiltrations were classified as type 1 and included. Type 1 CNVs accompanied by intraretinal fluid were also excluded because intraretinal layer thickness measurements are unclear in the presence of intraretinal fluid. Patients with thick choroid type 1 CNV such as aneurysmal type 1 CNV or polypoidal choroidal vasculopathy (PCV), which has a high prevalence in Asia, were included.
Overall, 94 eyes of 94 consecutive patients that met the inclusion criteria were enrolled in the study. Additionally, 35 normal eyes of 35 consecutive study participants who presented for mild cataracts or floaters and who did not have any morphological/functional retinal disorders were included as the control group. Based on fibrovascular PED identified by OCT, leakage and/or staining identified by FA, and the presence of CNV identified by ICGA, the relative location of the CNV to the foveal center was documented as subfoveal (CNV located under the foveal center), juxtafoveal (CNV located not under the foveal center but within the 1 mm foveal diameter circle), or extrafoveal (CNV located outside of the 1 mm foveal diameter circle). If a discrepancy existed among the three modalities, the CNV location was graded as the majority of modalities determined.
After a diagnosis of nAMD with type 1 CNV, the investigated patients were treated with an initial loading dose of aflibercept given via intravitreal injection every 3 months. After the loading period, patients were treated in bi- or trimonthly intervals as required by their clinical needs, following the treat & extend strategy13,14. Patients who had residual or increasing SRF with declining visual acuity on bimonthly aflibercept were switched to monthly ranibizumab (Lucentis; Genentech, San Francisco, CA, USA) or bevacizumab (Avastin; Genentech) at the ophthalmologist’s discretion. BCVA and foveal OCT data at baseline, after three loading doses of aflibercept, and at the 1-year follow-up were gathered and analyzed. The information regarding the total number of injections and follow-up duration was also documented.
The OCT biomarkers analyzed in the current study included: foveal ONL thickness, foveal SRF height, subfoveal choroidal thickness, and the presence of the foveal EZ. SRF height and subfoveal choroidal thickness were chosen because these biomarkers can be associated with the activity and vascularity of CNV. The data were produced with spectral domain (SD)-OCT in the enhanced depth imaging mode (Spectralis OCT; Heidelberg Engineering, Heidelberg, Germany). All OCT parameters were measured manually by using the computer-based caliper measurement tool in the SD-OCT system. On-axis images were selected and analyzed when several OCT images were available. The foveal ONL thickness was defined as the distance between the internal and external limiting membranes at the center of the fovea15,16. The foveal SRF height was measured from the tip of the RPE layer to the outer border of the detached retina at the fovea17. Subfoveal choroidal thickness was defined as the vertical distance from the outer border of the RPE to the inner border of the sclera at the fovea18. All of the aforementioned measurements were performed by two operators (EJS and SL) according to the same protocol without any knowledge of the other measurements. Intergrader agreement, measured with the intraclass correlation coefficient (ICC), was excellent between the two graders (see Supplementary Table S1). The mean of the two measurements was used in the analysis.
The presence of EZ was graded in a single horizontal scan through the fovea. EZ was defined as a hyperreflective line between the external limiting membrane and RPE, which is also hyperreflective in OCT. The presence of EZ was graded at baseline, after three monthly loading doses of aflibercept, and at the 1-year follow-up. Each operator (EJS and SL) graded without the knowledge of the other’s grading. In case of a discrepancy between the two graders, a third blinded grader (KTK) adjudicated the decision. The inter-rater reliability was estimated with Cohen's kappa coefficient.
Baseline BCVA and OCT parameters were compared between the study and control groups. Changes in OCT parameters during treatment and their relationship to the final visual outcome were analyzed. The study group was subsequently divided into two subgroups, based on the difference in foveal ONL thickness at baseline and after three loading doses of aflibercept: the poor response group (group A) and the good response group (group B). To divide the overall study participants into 50% of good responders and 50% of poor responders, the median change in foveal ONL thickness in the overall study population was used as the cut-off value, which was + 10 µm. Hence, the 94 eyes were categorized into 46 eyes for the poor responders (group A) and 48 eyes for the good responders (group B). BCVA and OCT parameters were compared between these groups.
SPSS version 21.0 (SPSS Inc, Chicago, IL, USA) was used for the statistical analysis. The independent t-test was used to compare variables between the study and control groups and between the subgroups. Chi-square analysis was used to compare EZ, age, and sex. Fisher’s exact test was used to compare the ratio of the CNV locations between subgroups. Multivariate linear regression was used to investigate the parameters that could predict the final visual outcome. In all the analyses, a value of p < 0.05 was considered as statistically significant.